Pricing
Docs
Blog
About
Contact

🎉 Hy4-preview is available on SiliconFlow. Try it NOW.

🎉 Hy4-preview is available on SiliconFlow. Try it NOW.

Blog

Latest articles, tutorials, and insights about AI infrastructure and model deployment.

Long Context vs. RAG: Which Should You Use for Large Documents?

Long context sends every document with the request. RAG sends the passages retrieval ranked highest. Compare token cost, updates, citations, and access control. ...

GLM-5.3-Flash vs. GLM-5.2 vs. Claude Opus 4.8: A Hands-On Coding and Agent Benchmark

GLM-5.3-Flash scores 63.4 on DeepSWE v1.1 against 46.2 for GLM-5.2, and lands within a point of Claude Opus 4.8 on Terminal-Bench. See the costs behind them. ...

GLM-5.3-Flash on SiliconFlow: Architecture, Pricing and Multimodal Limits

A 320B mixture-of-experts model with 18B active, 1M context, and text, image, video, and file input. Here is what one million tokens actually costs. ...

DeepSeek V4 Pro on SiliconFlow: Reasoning Levels, Pricing and Long-Task Cost

Low handles routine edits, High covers most multi-step work, and Max earns its cost on ambiguous failures. Route by task risk, not by habit. ...

What Is LLM Inference? Latency, Throughput, Batching, and Cost Explained

Prefill sets time to first token and decode sets the rest. See how TTFT, TPOT, batching, GPU memory, and serving choices change latency and cost. ...

What Is an AI Gateway? How It Manages Multiple LLM APIs

An AI gateway sits between your app and every model call. Compare it with a traditional API gateway, and see how routing, fallback, and quotas work. ...

RAG vs. Fine-Tuning: Which Should You Use for LLM Applications?

RAG adds outside knowledge at request time; fine-tuning changes how the model behaves. Compare data needs, freshness, latency, and cost before you pick. ...

Serverless Inference vs Reserved GPUs: Which Is Better for Your AI Workload?

Serverless inference bills per model call; reserved GPUs trade that flexibility for predictable capacity. Compare cost, scaling, and operational load. ...

How to Use an AI Video Generation API: Submit Jobs, Poll Status, and Store Results

Video generation outlasts a single HTTP request. See how requestId polling, bounded backoff, and object storage keep a long-running job reliable. ...

Do You Need a Reranker for RAG? Embeddings, Reranking, and Two-Stage Retrieval Explained

Most RAG failures are ranking problems, not retrieval gaps. Compare recall@k and nDCG, then decide whether a reranker earns its latency and cost. ...

SiliconFlow API Rate Limits and 429 Retry Strategies

SiliconFlow rate limits apply per account and per model, so bursts, shared keys or retries trigger 429 errors. Backoff, jitter and queues keep traffic stable. ...

OpenAI SDK Setup for SiliconFlow API Integration

Set the SiliconFlow base URL and model ID in the OpenAI SDK, then verify streaming, tools, vision, reasoning, limits, and errors before production. ...

Tencent Hy4 Preview: 1M Context, Coding, and API Guide

Evaluate Tencent Hy4 Preview’s 1M context, MoE architecture, coding benchmarks, deployment options, preview limits, and API cost factors in production. ...

DeepSeek V4.1 Flash: Architecture, Benchmarks, Pricing, and API Migration

Compare DeepSeek V4.1 Flash architecture, multimodal benchmarks, direct API pricing, migration steps, and cost controls for coding agents in production. ...

DeepSeek V4 Pro Function Calling: A Python Tutorial for Tool-Using Agents

DeepSeek V4 Pro Function Calling: A Python Tutorial for Tool-Using Agents

Build a DeepSeek V4 Pro tool-calling agent in Python. Validate function arguments, execute approved tools, return results, and handle errors safely. ...

How to Use DeepSeek V4 Pro in Claude Code: SiliconFlow Setup and Troubleshooting

How to Use DeepSeek V4 Pro in Claude Code: SiliconFlow Setup and Troubleshooting

Connect Claude Code to DeepSeek V4 Pro through SiliconFlow by setting the Anthropic base URL, mapping the model ID, and testing the Messages endpoint. ...

GLM-5.3 vs. GLM-5.3-Flash: Coding, Multimodal Input, and API Cost Compared

GLM-5.3 vs. GLM-5.3-Flash: Coding, Multimodal Input, and API Cost Compared

GLM-5.3 and GLM-5.3-Flash differ in coding strength, image input, and token pricing. Compare context limits and API cost before routing work between them. ...

DeepSeek V4 Pro vs. GLM 5.3: Which Should You Choose for Coding and Agents?

DeepSeek V4 Pro vs. GLM 5.3: Which Should You Choose for Coding and Agents?

DeepSeek V4 Pro costs less per token, while GLM 5.3 leads several long-horizon coding benchmarks. Compare pricing, context, and agent fit to choose. ...

GLM 5.3 API Guide: Pricing, Model ID, and Your First Request

GLM 5.3 API Guide: Pricing, Model ID, and Your First Request on SiliconFlow

GLM-5.3 on SiliconFlow costs $1.40 per 1M input tokens and $4.40 per 1M output tokens. See the model ID, a Python first request, streaming, and error fixes. ...

GLM-5.3 Now Live on SiliconFlow

A New Flagship Model for Complex Software Engineering and Long-Horizon Agent Tasks ...

Cover image for GLM-5.3 coding guide showing use cases, benchmarks, and limits for software engineering tasks

Is GLM-5.3 Good for Coding? Best Use Cases and Current Limits

GLM-5.3 improves terminal coding and long-horizon agent tasks. Review its strengths, benchmarks, practical limits, and SiliconFlow testing steps. ...

GLM-5.3 vs Claude Fable 5 coding benchmarks, API pricing, and context window comparison on SiliconFlow

GLM-5.3 vs Claude Fable 5 Coding Benchmarks and Cost

GLM-5.3 and Claude Fable 5 both target coding agents, but GLM-5.3 costs far less per token. Compare benchmarks, context windows, and API pricing to decide. ...

DeepSeek V4 Flash for Coding Agents — cover image showing AI coding assistant with code and lightning bolt

DeepSeek V4 Flash Coding Agent Use Cases and Pricing

DeepSeek V4 Flash excels at routine coding agent tasks. Route work based on validation results, task ambiguity, and cost per successful completion. ...

DeepSeek-V4-Pro-0813 Now Live on SiliconFlow

Enhanced Agent Capabilities for Production Workflows ...

FP8 Inference: What It Means for Your API Bill

FP8 can improve inference efficiency, but precision alone does not set API prices. Compare token rates, caching, retries, and cost per successful task. ...

Prompt Caching on SiliconFlow: How Cache Read Pricing Can Reduce Your API Bill

See how SiliconFlow cache read pricing changes API cost estimates, where prompt caching saves the most, and which workload factors affect real savings. ...

Best Cheap API for SillyTavern: DeepSeek-V4-Flash, GLM-5.2, and Long-Context Cost

Compare DeepSeek-V4-Flash and GLM-5.2 costs for long SillyTavern sessions, including context growth, token pricing, setup, and practical cost examples. ...

DeepSeek V4 Flash 0731 Now Live on SiliconFlow

The official release of DeepSeek V4 Flash, same size and architecture as the preview, re-post-trained for a significant jump in agentic capability. ...

Best API models for Cline compared across DeepSeek V4 Flash, DeepSeek V4 Pro, and GLM 5.2 for coding cost and performance

Models for Cline: DeepSeek V4 Flash, DeepSeek V4 Pro, and GLM 5.2 Cost Compared

Compare DeepSeek V4 Flash, DeepSeek V4 Pro, and GLM 5.2 as Cline API models, with cost examples and routing guidance for coding workflows. ...

Claude Opus alternative comparison showing GLM-5.2, DeepSeek-V4-Pro, DeepSeek-V4-Flash, Kimi-K2.6, and DeepSeek-V3.2 for coding workflows

Claude Opus Alternative for Coding: GLM-5.2 and DeepSeek-V4 Compared

Compare GLM-5.2, DeepSeek-V4-Pro, DeepSeek-V4-Flash, Kimi-K2.6, and DeepSeek-V3.2 as Claude Opus coding alternatives with pricing and context. ...

Claude Code alternatives comparison covering CLI agents, IDE assistants, open-source coding tools, and API-based developer workflows

Claude Code Alternatives: Open-Source, CLI, IDE, and API-Based Coding Tools Compared

Compare Claude Code alternatives across CLI, IDE, open-source, and API-based coding tools, and see how SiliconFlow powers flexible model access. ...

DeepSeek V4 API pricing comparison for Pro and Flash models with context window, cached input, and cost examples

DeepSeek V4 API Pricing: Pro vs Flash, Context Window, and Cost Examples

Compare DeepSeek V4 Pro and Flash API pricing, context window, and cached-input costs with worked examples for chatbots, RAG, and coding agents. ...

How to use Open Design with SiliconFlow's APIs

Use SiliconFlow as the built-in model provider in Open Design — the open-source, agent-native design workspace — to generate prototypes, landing pages, dashboards, decks, and more with 200+ models. ...

SiliconFlow-style cover image for Kimi model comparison guide

Kimi K3 vs. Kimi K2.7 Code vs. Kimi K2.6: Which Model Should You Choose?

Compare Kimi K3, Kimi K2.7 Code, and Kimi K2.6 by context length, coding fit, JSON support, multimodal use, latency, and task cost. ...

SiliconFlow-style cover image for Kimi K3 API on SiliconFlow

Kimi K3 on SiliconFlow: Features, Pricing, and API Access

Learn Kimi K3 features, SiliconFlow API access, pricing factors, context limits, JSON Mode, tool calling, and production planning checks. ...

Kimi K3 Now Live on SiliconFlow: The First Open 3T-Class Model at Frontier-Level Performance

Kimi K3 is now available on SiliconFlow — the first open 3T-class model with 2.8T parameters, 1M-token context, and native vision. OpenAI- and Anthropic-compatible API, starting at $3.0/M input tokens. Integrates with Claude Code, Cline, Cursor, and more. ...

Developer reviewing the Hy3 AI model dashboard with Mixture-of-Experts architecture, 256K context, and coding workflow tools

What Is Hy3? Tencent Hunyuan’s MoE Model Explained for Developers

Tencent's Hy3 MoE model leverages 295B total and 21B active parameters to power reasoning, coding, and agent workflows. See how its sparse architecture works. ...

Developer using a laptop and large monitor to build long-context AI agent workflows with code, document analysis, and API tools

Hy3 API Guide: How to Use Tencent Hunyuan Hy3 on SiliconFlow

Integrate the Tencent Hunyuan Hy3 API seamlessly for coding agents and long-context tasks. Claim your two weeks of free OpenAI-compatible access today. ...

Hy3 Now Live on SiliconFlow: Advancing Agent Capabilities, Ready for Real-World Workflows

Hy3, Tencent Hy Team's next-generation flagship, is now live on SiliconFlow. Free fro the first 2 weeks. ...

LongCat-2.0 Now on SiliconFlow: 1.6T MoE, Native 1M Context, Built for Agentic Coding

LongCat-2.0 is now live on SiliconFlow Meituan LongCat's next-generation trillion-parameter flagship model, and it's the full model behind "Owl Alpha", the anonymous model that reached Top 3 globally by daily volume on OpenRouter. ...

SiliconFlow Summer Rush丨GLM 5.2 Week : The More GLM 5.2 You Run, the More You Win

SiliconFlow Summer Rush — GLM 5.2 Week is Live ...

Team discussing GLM-5.2 vs GPT-5.5 in a modern office, with large screens showing AI model pricing, context window comparison, API access, and deployment options.

GLM-5.2 vs GPT-5.5: Pricing, Context Window & Open-Source Access

GLM-5.2 costs 77% to 87% less than GPT-5.5 for large workloads while offering an identical 1M token context window and MIT-licensed open weights. ...

Developer reviewing an AI model API dashboard with workflow, token flow, pricing analysis, and long context visualization on a large monitor in a modern office.

GLM 5.2 API Guide: GLM-5.2 Pricing, Model ID, 1M Context, and Cost Examples

Optimize your 1M context window costs. Learn how GLM-5.2 prompt caching reduces your SiliconFlow Serverless API bills with real examples. ...

Kimi K2.7 Code Now on SiliconFlow: Better Agentic Coding, 30% Fewer Thinking Tokens

Kimi K2.7 Code is Moonshot AI's latest open-source, coding-focused agentic model, now live on SiliconFlow. ...

GLM-5.2 Now on SiliconFlow: 1M Context, Long-Horizon Engineering, Near-Opus Coding

In this article, we have tested and compared GLM-5.2 with Opus 4.8, GPT-5.5, and GLM-5.1 under the same conditions, highlighting where GLM-5.2 reaches frontier-level performance while offering a major cost advantage on SiliconFlow. ...

Nex-N2-Pro Now on SiliconFlow: Agentic Thinking, Frontier Coding, and Long-Horizon Execution

Nex-N2-Pro is an agentic model with Agentic Thinking. Built for real-world productivity, it turns reasoning into executable, verifiable, and iterable action across agentic coding, deep research, tool calling, and terminal execution. ...

How to Use CodeWhale with SiliconFlow: DeepSeek V4 Terminal Coding Agent Setup

CodeWhale brings DeepSeek V4 to your terminal with a structured harness and recursive improvement that prevents model drift. With SiliconFlow as a native provider, setup takes a single command. ...

MiniMax M3 Now on SiliconFlow: Frontier Coding, 1M Context, and Native Multimodality

MiniMax M3, the first open-weight model to bring frontier coding, million-token context, and native multimodality together in one system, is now officially live on SiliconFlow. ...

How to Build a Personal Wiki Knowledge Base: Andrej Karpathy's Pattern with SiliconFlow & OpenCode

Build an automated wiki knowledge base based on Andrej Karpathy's llm-wiki sharing. Crawl, analyze, and compile with SiliconFlow, OpenCode, and oh-my-openagent. ...

How to Deploy an AI Assistant on Discord Using Hermes Agent

This guide takes you further, from choosing the right model on SiliconFlow to building your daily assistant on Discord. Plus, a set of practical tips and best practices to help you get more out of every session. ...

How to use CC Switch with SiliconFlow APIs

This guide will walk you through how to connect SiliconFlow to CC Switch and start powering your agentic workflows with 200+ frontier models in just a few steps. ...

How to use Continue in Vs Code with SiliconFlow APIs

Use SiliconFlow's frontier open-source models in Continue for VS Code. Step-by-step guide to configure DeepSeek V4, GLM-5.1, Tab autocomplete, and more. ...

Qwen3.6 Series Now on SiliconFlow: Flagship-Level Coding in Compact Models

Built on community feedback, Qwen3.6 delivers substantial upgrades in coding agents, agentic workflows, and multimodal understanding with native thinking capabilities. ...

Qwen3.5 Series Now on SiliconFlow: Multimodal AI Excellence

The Qwen3.5 series, Alibaba's latest advancement in foundation model development, represents a significant leap forward in multimodal AI, featuring five powerful models ranging from 9B to 397B parameters. ...

Gemma 4 Now on SiliconFlow: Frontier Multimodal Intelligence

The Gemma 4 family of multimodal models by Google DeepMind delivers frontier-level performance, built for advanced reasoning and agentic workflows, and brings comprehensive multimodal capabilities and exceptional performance across diverse benchmarks. ...

DeepSeek-V4 Now on SiliconFlow: Million-Token Context Intelligence

DeepSeek-V4 introduces two powerful MoE models with groundbreaking 1M-token context windows, seting a new benchmark for unprecedented long-context efficiency, advanced reasoning capabilities, and SOTA performance across coding, mathematics, and agentic tasks. ...

Kimi K2.6 Now on SiliconFlow: SOTA Long-horizon Coding

Kimi K2.6, Moonshot AI's latest open-source multimodal agentic model. This release brings exceptional long-horizon coding capabilities, proactive autonomous execution, and swarm-based task orchestration, enabling developers to achieve production-grade agentic workflows with greater efficiency and reliability. ...

How to use Roo Code with SiliconFlow APIs

In this guide, we'll walk you through how to intergrate SiliconFlow APIs into Roo Code, giving you access to 200+ frontier models at competitive pricing. ...

How to use Cline with SiliconFlow APIs to Build Your rojects

In this guide, we'll show you how to integrate SiliconFlow's APIs into Cline, giving you access to frontier coding models like GLM-5.1, MiniMax M2.5, and Kimi K2.5 at cost-effective pricing. ...

How to use Chub AI with SiliconFlow's APIs

In this guide, we'll show you how to integrate SiliconFlow APIs into Chub AI, giving you access to 200+ top-tier models to power your RP experience. ...

GLM-5.1 Now on SiliconFlow: Autonomous Engineering at Scale

GLM-5.1, Z.AI's latest flagship model designed for long-horizon agentic engineering, achieves SOTA performance on SWE-Bench Pro and can work autonomously on complex tasks for up to 8 hours. ...

How to use Janitor AI with SiliconFlow APIs

In this guide, we'll show you how to use SiliconFlow APIs in Janitor AI, giving you access to 200+ open-source and frontier models at competitive pricing. ...

GLM-5V-Turbo Now on SiliconFlow: Vision Coding Redefined

GLM-5V-Turbo, Z. AI's multimodal coding foundation model built for vision-based coding and agent-driven tasks, brings native multimodal understanding and balanced visual-programming capabilities. ...

How to use Hermes Agent with SiliconFlow APIs

In this guide, we'll walk you through how to connect SiliconFlow's APIs to Hermes Agent, so you can power your agent with 200+ SOTA AI models in just a few steps. ...

MiniMax M2.5 Now on SiliconFlow: Built For Real-World Productivity

Discover MiniMax M2.5 — a frontier agentic model with SOTA coding, tool use, and office productivity, now accessible on SiliconFlow. ...

Step 3.5 Flash Now on SiliconFlow: Deep Thinking at Flash Speed

Step 3.5 Flash, the most capable StepFun AI's open-source foundation model is here. It is engineered to deliver frontier reasoning and agentic capabilities with exceptional efficiency. ...

GLM-5 Now on SiliconFlow: SOTA Open-Source Model Built for Agentic Engineering

Engineered for the era of agentic engineering, GLM-5 moves beyond 'vibe coding' to deliver reliable performance for complex systems and long-horizon tasks. ...

Kimi K2.5 Now on SiliconFlow: SOTA on Visual Agentic Intelligence

Built on Kimi K2 through continued pretraining over approximately 15T mixed visual and text tokens, it delivers state-of-the-art coding and vision capabilities as a native multimodal model. ...

MiniMax M2.1 on SiliconFlow: Enhanced Multi-Language Programming, Built for Real-World Complex Tasks

MiniMax M2.1, Open-source MoE model (10B active/230B total) built for real-world development—excels at multi-language programming, office automation, and agent workflows. ...

GLM-4.7 Now on SiliconFlow: Advanced Coding, Reasoning & Tool Use Capabilities

We're excited to announce that GLM-4.7, Z.ai's latest flagship model, is now available on SiliconFlow with Day 0 support. ...

FLUX.2 Now on SiliconFlow: Designed for Real-world Creative Workflows

We are excited to bring FLUX.2 [pro] and FLUX.2 [flex] to SiliconFlow, setting a new standard for real-world creative workflows. ...

GLM-4.6V Now on SiliconFlow: Native Multimodal Tool Use Meets SoTA Visual Intelligence

Z.ai's GLM-4.6V, featuring native function calling, 131K context, and SoTA visual understanding, is now available on SiliconFlow for building multimodal agents in real-world business scenarios. ...

Z-Image-Turbo Now on SiliconFlow: Photorealistic & Bilingual Text Rendering

Z-Image-Turbo, Alibaba Tongyi’s lightweight 6B text-to-image model, is now live on SiliconFlow, delivering photorealistic generation and bilingual text rendering on par with leading commercial models. ...

DeepSeek-V3.2 Now on SiliconFlow: Reasoning-first model built for agents

DeepSeek-V3.2 is now live on SiliconFlow, bringing GPT-5-level reasoning, efficient agentic performance, 164K context window and advanced tool-use to power your workflows. ...

Kimi K2 Thinking Now on SiliconFlow: Thinking Agent That Reasons and Acts

Designed as a thinking agent, it reasons step by step and can execute up to 200-300 sequential tool calls without human interference, thinking coherently across hundreds of steps to solve complex problems. ...

8 Key Insights on AI Infra from the co-founder of SiliconFlow

Pan Yang, co-founder of SiliconFlow, delivered a speech entitled “AI Infra: For Whom and Why?” at “Real-Time AI Infra Session” of Convo AI & RTE 2025. There are 8 core insights into the field of AI Infra. ...

MiniMax-M2 Now on SiliconFlow: Frontier-Style Coding and Agentic Intelligence

MiniMax-M2 is now live on SiliconFlow — a compact MoE model that delivers frontier-level reasoning and coding performance with efficiency and affordability. ...

Qwen3-VL-32B Now on SiliconFlow: Flagship-Level Intelligence with Dense-Model Efficiency

Qwen3-VL-32B is now live on SiliconFlow — a dense 32B model delivering flagship-level multimodal reasoning with faster responses, lower costs, and exceptional efficiency. ...

Qwen3-VL-8B Now on SiliconFlow: Small Model, Big Vision

Qwen3-VL-8B — the newest member of the Qwen3-VL family — is now live on SiliconFlow, offering full-scale multimodal reasoning at just 8B parameters. ...

Unlocking the Future of Video Technology: Introducing Hunyuan Video by Tencent

As video becomes central to communication and creativity, Tencent is reshaping the field with Hunyuan Video — an open-source AI platform built to revolutionize how we create, interact with, and share video content. ...

Zoom’s Transformation: Becoming an AI-First Company for the Future of Work

Work has evolved, and so has Zoom — now an AI-first company focused on smarter, faster collaboration beyond video calls, offering integrated tools that power the future of hybrid and remote work. ...

Ring-1T Now on SiliconFlow

Ring-1T Now on SiliconFlow: The World's First Open-Source Trillion-Parameter Thinking Model

Today, we're excited to bring Ring-1T to SiliconFlow — the first open-source trillion-parameter thinking model developed by Ant Group's inclusionAI team. ...

Ling-1T Now Live on SiliconFlow

Ling-1T Now Live on SiliconFlow: A Trillion-Scale Leap in Efficient Reasoning

Ling-1T — the first flagship model in the Ling 2.0 series — is now live on SiliconFlow, delivering state-of-the-art reasoning performance with an exceptional balance of accuracy, speed, and cost-efficiency. ...

Qwen3-VL on SiliconFlow

Qwen3-VL on SiliconFlow: Next-Gen VLM with Better World Understanding

Qwen3-VL, the most powerful vision-language model in the Qwen series, is now live on SiliconFlow — delivering advanced multimodal reasoning, 262K context, 32-language OCR, and up to 235B parameters for next-level AI understanding. ...

DeepSeek-V3.2-Exp Now on SiliconFlow

DeepSeek-V3.2-Exp Now on SiliconFlow: Efficient Long-Context Reasoning at 50% Lower Cost

DeepSeek-V3.2-Exp is now live on SiliconFlow — delivering faster, more efficient long-context reasoning and coding with DeepSeek Sparse Attention and a 164K context window. ...

Qwen3-Omni Now on SiliconFlow

Qwen3-Omni Now on SiliconFlow: Alibaba's Next-Gen Multimodal Foundation Model

Today, we're launching Qwen3-Omni on SiliconFlow — a native omni-modal foundation model that processes text, vision, audio, and video inputs with real-time streaming output, redefining multimodal AI capabilities. ...

GLM-4.6 Now on SiliconFlow: Advanced Agentic, Reasoning and Coding Capabilities

SiliconFlow is excited to bring you GLM-4.6, the latest version of Z.ai’s flagship model — featuring enhanced long-context reasoning, deeper agent integration, and improved coding efficiency. ...

Hunyuan-MT-7B Now On SiliconFlow

Hunyuan-MT-7B Now On SiliconFlow: High-quality Translation Model Supporting 33 Languages

SiliconFlow is thrilled to bring Hunyuan-MT-7B to our model catalog — Tencent's first open-source multilingual translation model supporting bidirectional translation across 33 major languages. ...

Ling-flash-2.0 Now on SiliconFlow

Ling-flash-2.0 Now on SiliconFlow: Flagship MoE Model Delivering SOTA Reasoning and High Efficiency

SiliconFlow is excited to bring you Ling-flash-2.0, the third MoE model under Ling 2.0 architecture, featuring a step forward in combining efficiency and reasoning ability. ...

Qwen-Image Is Here on SiliconFlow

Qwen-Image Is Here on SiliconFlow: Superior Text Rendering & Precise Image Editing

Today, Qwen-Image — a 20B MMDiT foundation model — is now available on SiliconFlow, offering advanced text rendering, precise editing, and versatile image generation. ...

Qwen-Image-Edit Now Available on SiliconFlow

Qwen-Image-Edit Now Available on SiliconFlow: Advanced Text, Semantic and Appearance Editing

Qwen-Image-Edit is now available on SiliconFlow. Powered by the Qwen-Image 20B model, it offers precise text editing and comprehensive semantic and appearance editing while preserving realism. ...

Ling-mini-2.0 Now on SiliconFlow

Ling-mini-2.0 Now on SiliconFlow: MoE Model with SOTA Performance & High Efficiency

SiliconFlow is excited to introduce Ling-mini-2.0 — Ant Group inclusionAI's MoE-based language model that redefines how efficient AI models can be. ...

Kimi K2-0905 Now on SiliconFlow

Kimi K2-0905 Now on SiliconFlow: Enhanced Coding Capabilities with Smoother Claude Code Integration

SiliconFlow is excited to welcome Kimi K2-0905 — Moonshot AI's most capable coding-focused upgrade that transforms how developers build and innovate. ...

Seed-OSS-36B-Instruct Now Available On SiliconFlow

Seed-OSS-36B-Instruct Now Available On SiliconFlow: Smarter AI That Thinks On-Demand

SiliconFlow is excited to bring Seed-OSS-36B-Instruct to our model catalog - ByteDance's revolutionary open-source model that puts AI reasoning control in your hands. ...

DeepSeek-V3.1 on SiliconFlow

DeepSeek-V3.1 on SiliconFlow: Hybrid Thinking, Smarter Tools and 164K Context Window

SiliconFlow is thrilled to bring DeepSeek-V3.1 to our model catalog — the latest upgrade from DeepSeek that takes AI one step closer to the Agent era. ...

penAI's gpt-oss Now Live on SiliconFlow

OpenAI's gpt-oss Now Live on SiliconFlow: Designed for Agentic Workflows, Advanced Reasoning and Tool Use

SiliconFlow is excited to announce the launch of gpt-oss-120B and gpt-oss-20B — state-of-the-art open-weight language models now available on our platform. ...

Wan 2.2 Series Now Available on SiliconFlow

Wan 2.2 Series Now Available on SiliconFlow：Generate Stable, Realistic and Cinematic Videos

SiliconFlow is excited to bring the Wan 2.2 series to our platform — a major upgrade to Wan's visual generative models, featuring significantly improved performance and superior visual quality. ...

GLM-4.5V: The World's Leading Open-Source Vision Reasoning Model Now on SiliconFlow

Today, we are excited to introduce that GLM-4.5V — the world’s best-performing open-source 100B-scale vision reasoning model — is now available on SiliconFlow. ...

Step3 Now Live on SiliconFlow: The Leading Open-source Multimodal Reasoning Model

Step3 Now Live on SiliconFlow: The Leading Open-source Multimodal Reasoning Model

Step3, Stepfun's latest cutting-edge multimodal reasoning model is now available on SiliconFlow. ...

Think Deeper, Act Faster: Qwen3-235B-A22B-Thinking-2507 Now Available on SiliconFlow

Think Deeper, Act Faster: Qwen3-235B-A22B-Thinking-2507 Now Available on SiliconFlow

With Qwen3-235B-A22B-Instruct-2507 (Non-Thinking mode) already demonstrating exceptional performance on SiliconFlow ...

GLM-4.5 Now Available on SiliconFlow

GLM-4.5 Now Available on SiliconFlow: Open-Source SOTA Model for Reasoning, Code, and Agentic Applications

Today, we're excited to integrate GLM-4.5 and GLM-4.5-Air, Z.ai's latest flagship model serie, into SiliconFlow platform. ...

Qwen3-235B-A22B-Instruct-2507: Leading Open-Source Model Now Live on SiliconFlow

Qwen3-235B-A22B-Instruct-2507: Leading Open-Source Model Now Live on SiliconFlow

Qwen has just released Qwen3-235B-A22B-Instruct-2507, an upgraded version of its flagship model Qwen3-235B-A22B Non-thinking. ...

Next-Gen Image Generation & Editing: FLUX.1 Kontext Now Available on SiliconFlow

Next-Gen Image Generation & Editing: FLUX.1 Kontext Now Available on SiliconFlow

SiliconFlow today introduces FLUX.1 Kontext [pro] and FLUX.1 Kontext [max], a suite of generative flow matching models that allows you to generate and edit images at the next level. ...

Kimi-K2 on SiliconFlow: Tailored for AI Agents, Priced to Scale

Kimi-K2 on SiliconFlow: Tailored for AI Agents, Priced to Scale

MoonShot AI's powerful open-source Kimi K2 Mixture-of-Experts (MoE) model is now available on SiliconFlow. ...

Run ERNIE-4.5-300B-A47B with SiliconFlow's API

The Baidu ERNIE team has announced the release of its ERNIE-4.5-300B-A47B, a powerful open-source large language model now available on the SiliconFlow platform. ...

Hunyuan-A13B-Instruct Now Available on SiliconFlow

The Tencent Hunyuan AI team announced the release of Hunyuan-A13B-Instruct, an open-source large language model (LLM) now available on the SiliconFlow platform. ...

FLUX.1 Kontext Dev on SiliconFlow：Redefining Ultra-Fast, Context-Aware Image Editing

FLUX.1 Kontext Dev on SiliconFlow：Redefining Ultra-Fast, Context-Aware Image Editing

FLUX.1 Kontext Dev, the world’s fastest and best context-aware open image editing model by Black Forest Labs, is now officially available on SiliconFlow. ...

MiniMax-M1-80k（456B）, the world’s first open-source hybrid-attention model at scale, is now available on SiliconFlow.

MiniMax-M1-80k Now Available on SiliconFlow

MiniMax-M1-80k（456B）, the world’s first open-source hybrid-attention model at scale, is now available on SiliconFlow. ...

DeepSeek-R1-0528 on SiliconFlow

Enhanced throughput, reduced hallucination, and improved reasoning — built for high-performance, high-reliability generative AI tasks. ...

Fast, Affordable, Powerful: Run Wan2.1 on SiliconFlow

Wan2.1, a comprehensive and open suite of video foundation models that pushes the boundaries of video generation, is now available on SiliconFlow. ...

Revolutionizing Spatial AI: Introducing World Labs’ 3D Generation Model

Generative AI has long focused on creating 2D content, but World Labs, co-founded by Fei-Fei Li, is taking a quantum leap forward — introducing a 3D generation model that turns a single image into an immersive, explorable world. ...

Unlocking Learning Potential: A Guide for Students on AI-Powered Academic Tools

Artificial intelligence is transforming education today—empowering students with personalized, efficient tools that enhance learning, boost productivity, and redefine how they achieve academic success. ...

DeepSeek-V3-0324 Is Now Live on SiliconFlow

DeepSeek-V3-0324 (671B) is now live on SiliconFlow — delivering major improvements in reasoning, writing, and math performance. ...

Mastering Creativity with Precision: How ControlNet Transforms Text-to-Image AI

AI-generated imagery is advancing rapidly, and ControlNet stands out as a breakthrough innovation — offering creators unprecedented precision and control in transforming text descriptions into highly customized visuals. ...

The AI Revolution in Advertising

The AI Revolution in Advertising: Transforming Marketing in the Digital Age

AI's rapid evolution is transforming advertising, as seen in Curious Refuge's first AI Advertising Competition, which showcased how AI-generated commercials are redefining creativity and marketing. ...

Introducing QwQ 32B-preview: A Breakthrough in Reasoning AI

In the fast-moving world of AI, Alibaba Cloud’s Qwen Team has introduced QwQ, an open-source experimental model pioneering human-like reasoning and redefining analytical intelligence. ...

Long Context vs. RAG: Which Should You Use for Large Documents?

Long context sends every document with the request. RAG sends the passages retrieval ranked highest. Compare token cost, updates, citations, and access control. ...

GLM-5.3-Flash vs. GLM-5.2 vs. Claude Opus 4.8: A Hands-On Coding and Agent Benchmark

GLM-5.3-Flash scores 63.4 on DeepSWE v1.1 against 46.2 for GLM-5.2, and lands within a point of Claude Opus 4.8 on Terminal-Bench. See the costs behind them. ...

GLM-5.3-Flash on SiliconFlow: Architecture, Pricing and Multimodal Limits

A 320B mixture-of-experts model with 18B active, 1M context, and text, image, video, and file input. Here is what one million tokens actually costs. ...

DeepSeek V4 Pro on SiliconFlow: Reasoning Levels, Pricing and Long-Task Cost

Low handles routine edits, High covers most multi-step work, and Max earns its cost on ambiguous failures. Route by task risk, not by habit. ...

What Is LLM Inference? Latency, Throughput, Batching, and Cost Explained

Prefill sets time to first token and decode sets the rest. See how TTFT, TPOT, batching, GPU memory, and serving choices change latency and cost. ...

What Is an AI Gateway? How It Manages Multiple LLM APIs

An AI gateway sits between your app and every model call. Compare it with a traditional API gateway, and see how routing, fallback, and quotas work. ...

RAG vs. Fine-Tuning: Which Should You Use for LLM Applications?

RAG adds outside knowledge at request time; fine-tuning changes how the model behaves. Compare data needs, freshness, latency, and cost before you pick. ...

Serverless Inference vs Reserved GPUs: Which Is Better for Your AI Workload?

Serverless inference bills per model call; reserved GPUs trade that flexibility for predictable capacity. Compare cost, scaling, and operational load. ...

How to Use an AI Video Generation API: Submit Jobs, Poll Status, and Store Results

Video generation outlasts a single HTTP request. See how requestId polling, bounded backoff, and object storage keep a long-running job reliable. ...

Do You Need a Reranker for RAG? Embeddings, Reranking, and Two-Stage Retrieval Explained

Most RAG failures are ranking problems, not retrieval gaps. Compare recall@k and nDCG, then decide whether a reranker earns its latency and cost. ...

SiliconFlow API Rate Limits and 429 Retry Strategies

SiliconFlow rate limits apply per account and per model, so bursts, shared keys or retries trigger 429 errors. Backoff, jitter and queues keep traffic stable. ...

OpenAI SDK Setup for SiliconFlow API Integration

Set the SiliconFlow base URL and model ID in the OpenAI SDK, then verify streaming, tools, vision, reasoning, limits, and errors before production. ...

Tencent Hy4 Preview: 1M Context, Coding, and API Guide

Evaluate Tencent Hy4 Preview’s 1M context, MoE architecture, coding benchmarks, deployment options, preview limits, and API cost factors in production. ...

DeepSeek V4.1 Flash: Architecture, Benchmarks, Pricing, and API Migration

Compare DeepSeek V4.1 Flash architecture, multimodal benchmarks, direct API pricing, migration steps, and cost controls for coding agents in production. ...

DeepSeek V4 Pro Function Calling: A Python Tutorial for Tool-Using Agents

DeepSeek V4 Pro Function Calling: A Python Tutorial for Tool-Using Agents

Build a DeepSeek V4 Pro tool-calling agent in Python. Validate function arguments, execute approved tools, return results, and handle errors safely. ...

How to Use DeepSeek V4 Pro in Claude Code: SiliconFlow Setup and Troubleshooting

How to Use DeepSeek V4 Pro in Claude Code: SiliconFlow Setup and Troubleshooting

Connect Claude Code to DeepSeek V4 Pro through SiliconFlow by setting the Anthropic base URL, mapping the model ID, and testing the Messages endpoint. ...

GLM-5.3 vs. GLM-5.3-Flash: Coding, Multimodal Input, and API Cost Compared

GLM-5.3 vs. GLM-5.3-Flash: Coding, Multimodal Input, and API Cost Compared

GLM-5.3 and GLM-5.3-Flash differ in coding strength, image input, and token pricing. Compare context limits and API cost before routing work between them. ...

DeepSeek V4 Pro vs. GLM 5.3: Which Should You Choose for Coding and Agents?

DeepSeek V4 Pro vs. GLM 5.3: Which Should You Choose for Coding and Agents?

DeepSeek V4 Pro costs less per token, while GLM 5.3 leads several long-horizon coding benchmarks. Compare pricing, context, and agent fit to choose. ...

GLM 5.3 API Guide: Pricing, Model ID, and Your First Request

GLM 5.3 API Guide: Pricing, Model ID, and Your First Request on SiliconFlow

GLM-5.3 on SiliconFlow costs $1.40 per 1M input tokens and $4.40 per 1M output tokens. See the model ID, a Python first request, streaming, and error fixes. ...

GLM-5.3 Now Live on SiliconFlow

A New Flagship Model for Complex Software Engineering and Long-Horizon Agent Tasks ...

Cover image for GLM-5.3 coding guide showing use cases, benchmarks, and limits for software engineering tasks

Is GLM-5.3 Good for Coding? Best Use Cases and Current Limits

GLM-5.3 improves terminal coding and long-horizon agent tasks. Review its strengths, benchmarks, practical limits, and SiliconFlow testing steps. ...

GLM-5.3 vs Claude Fable 5 coding benchmarks, API pricing, and context window comparison on SiliconFlow

GLM-5.3 vs Claude Fable 5 Coding Benchmarks and Cost

GLM-5.3 and Claude Fable 5 both target coding agents, but GLM-5.3 costs far less per token. Compare benchmarks, context windows, and API pricing to decide. ...

DeepSeek V4 Flash for Coding Agents — cover image showing AI coding assistant with code and lightning bolt

DeepSeek V4 Flash Coding Agent Use Cases and Pricing

DeepSeek V4 Flash excels at routine coding agent tasks. Route work based on validation results, task ambiguity, and cost per successful completion. ...

DeepSeek-V4-Pro-0813 Now Live on SiliconFlow

Enhanced Agent Capabilities for Production Workflows ...

FP8 Inference: What It Means for Your API Bill

FP8 can improve inference efficiency, but precision alone does not set API prices. Compare token rates, caching, retries, and cost per successful task. ...

Prompt Caching on SiliconFlow: How Cache Read Pricing Can Reduce Your API Bill

See how SiliconFlow cache read pricing changes API cost estimates, where prompt caching saves the most, and which workload factors affect real savings. ...

Best Cheap API for SillyTavern: DeepSeek-V4-Flash, GLM-5.2, and Long-Context Cost

Compare DeepSeek-V4-Flash and GLM-5.2 costs for long SillyTavern sessions, including context growth, token pricing, setup, and practical cost examples. ...

DeepSeek V4 Flash 0731 Now Live on SiliconFlow

The official release of DeepSeek V4 Flash, same size and architecture as the preview, re-post-trained for a significant jump in agentic capability. ...

Best API models for Cline compared across DeepSeek V4 Flash, DeepSeek V4 Pro, and GLM 5.2 for coding cost and performance

Models for Cline: DeepSeek V4 Flash, DeepSeek V4 Pro, and GLM 5.2 Cost Compared

Compare DeepSeek V4 Flash, DeepSeek V4 Pro, and GLM 5.2 as Cline API models, with cost examples and routing guidance for coding workflows. ...

Claude Opus alternative comparison showing GLM-5.2, DeepSeek-V4-Pro, DeepSeek-V4-Flash, Kimi-K2.6, and DeepSeek-V3.2 for coding workflows

Claude Opus Alternative for Coding: GLM-5.2 and DeepSeek-V4 Compared

Compare GLM-5.2, DeepSeek-V4-Pro, DeepSeek-V4-Flash, Kimi-K2.6, and DeepSeek-V3.2 as Claude Opus coding alternatives with pricing and context. ...

Claude Code alternatives comparison covering CLI agents, IDE assistants, open-source coding tools, and API-based developer workflows

Claude Code Alternatives: Open-Source, CLI, IDE, and API-Based Coding Tools Compared

Compare Claude Code alternatives across CLI, IDE, open-source, and API-based coding tools, and see how SiliconFlow powers flexible model access. ...

DeepSeek V4 API pricing comparison for Pro and Flash models with context window, cached input, and cost examples

DeepSeek V4 API Pricing: Pro vs Flash, Context Window, and Cost Examples

Compare DeepSeek V4 Pro and Flash API pricing, context window, and cached-input costs with worked examples for chatbots, RAG, and coding agents. ...

How to use Open Design with SiliconFlow's APIs

Use SiliconFlow as the built-in model provider in Open Design — the open-source, agent-native design workspace — to generate prototypes, landing pages, dashboards, decks, and more with 200+ models. ...

SiliconFlow-style cover image for Kimi model comparison guide

Kimi K3 vs. Kimi K2.7 Code vs. Kimi K2.6: Which Model Should You Choose?

Compare Kimi K3, Kimi K2.7 Code, and Kimi K2.6 by context length, coding fit, JSON support, multimodal use, latency, and task cost. ...

SiliconFlow-style cover image for Kimi K3 API on SiliconFlow

Kimi K3 on SiliconFlow: Features, Pricing, and API Access

Learn Kimi K3 features, SiliconFlow API access, pricing factors, context limits, JSON Mode, tool calling, and production planning checks. ...

Kimi K3 Now Live on SiliconFlow: The First Open 3T-Class Model at Frontier-Level Performance

Kimi K3 is now available on SiliconFlow — the first open 3T-class model with 2.8T parameters, 1M-token context, and native vision. OpenAI- and Anthropic-compatible API, starting at $3.0/M input tokens. Integrates with Claude Code, Cline, Cursor, and more. ...

Developer reviewing the Hy3 AI model dashboard with Mixture-of-Experts architecture, 256K context, and coding workflow tools

What Is Hy3? Tencent Hunyuan’s MoE Model Explained for Developers

Tencent's Hy3 MoE model leverages 295B total and 21B active parameters to power reasoning, coding, and agent workflows. See how its sparse architecture works. ...

Developer using a laptop and large monitor to build long-context AI agent workflows with code, document analysis, and API tools

Hy3 API Guide: How to Use Tencent Hunyuan Hy3 on SiliconFlow

Integrate the Tencent Hunyuan Hy3 API seamlessly for coding agents and long-context tasks. Claim your two weeks of free OpenAI-compatible access today. ...

Hy3 Now Live on SiliconFlow: Advancing Agent Capabilities, Ready for Real-World Workflows

Hy3, Tencent Hy Team's next-generation flagship, is now live on SiliconFlow. Free fro the first 2 weeks. ...

LongCat-2.0 Now on SiliconFlow: 1.6T MoE, Native 1M Context, Built for Agentic Coding

LongCat-2.0 is now live on SiliconFlow Meituan LongCat's next-generation trillion-parameter flagship model, and it's the full model behind "Owl Alpha", the anonymous model that reached Top 3 globally by daily volume on OpenRouter. ...

SiliconFlow Summer Rush丨GLM 5.2 Week : The More GLM 5.2 You Run, the More You Win

SiliconFlow Summer Rush — GLM 5.2 Week is Live ...

Team discussing GLM-5.2 vs GPT-5.5 in a modern office, with large screens showing AI model pricing, context window comparison, API access, and deployment options.

GLM-5.2 vs GPT-5.5: Pricing, Context Window & Open-Source Access

GLM-5.2 costs 77% to 87% less than GPT-5.5 for large workloads while offering an identical 1M token context window and MIT-licensed open weights. ...

Developer reviewing an AI model API dashboard with workflow, token flow, pricing analysis, and long context visualization on a large monitor in a modern office.

GLM 5.2 API Guide: GLM-5.2 Pricing, Model ID, 1M Context, and Cost Examples

Optimize your 1M context window costs. Learn how GLM-5.2 prompt caching reduces your SiliconFlow Serverless API bills with real examples. ...

Kimi K2.7 Code Now on SiliconFlow: Better Agentic Coding, 30% Fewer Thinking Tokens

Kimi K2.7 Code is Moonshot AI's latest open-source, coding-focused agentic model, now live on SiliconFlow. ...

GLM-5.2 Now on SiliconFlow: 1M Context, Long-Horizon Engineering, Near-Opus Coding

In this article, we have tested and compared GLM-5.2 with Opus 4.8, GPT-5.5, and GLM-5.1 under the same conditions, highlighting where GLM-5.2 reaches frontier-level performance while offering a major cost advantage on SiliconFlow. ...

Nex-N2-Pro Now on SiliconFlow: Agentic Thinking, Frontier Coding, and Long-Horizon Execution

Nex-N2-Pro is an agentic model with Agentic Thinking. Built for real-world productivity, it turns reasoning into executable, verifiable, and iterable action across agentic coding, deep research, tool calling, and terminal execution. ...

How to Use CodeWhale with SiliconFlow: DeepSeek V4 Terminal Coding Agent Setup

CodeWhale brings DeepSeek V4 to your terminal with a structured harness and recursive improvement that prevents model drift. With SiliconFlow as a native provider, setup takes a single command. ...

MiniMax M3 Now on SiliconFlow: Frontier Coding, 1M Context, and Native Multimodality

MiniMax M3, the first open-weight model to bring frontier coding, million-token context, and native multimodality together in one system, is now officially live on SiliconFlow. ...

How to Build a Personal Wiki Knowledge Base: Andrej Karpathy's Pattern with SiliconFlow & OpenCode

Build an automated wiki knowledge base based on Andrej Karpathy's llm-wiki sharing. Crawl, analyze, and compile with SiliconFlow, OpenCode, and oh-my-openagent. ...

How to Deploy an AI Assistant on Discord Using Hermes Agent

This guide takes you further, from choosing the right model on SiliconFlow to building your daily assistant on Discord. Plus, a set of practical tips and best practices to help you get more out of every session. ...

How to use CC Switch with SiliconFlow APIs

This guide will walk you through how to connect SiliconFlow to CC Switch and start powering your agentic workflows with 200+ frontier models in just a few steps. ...

How to use Continue in Vs Code with SiliconFlow APIs

Use SiliconFlow's frontier open-source models in Continue for VS Code. Step-by-step guide to configure DeepSeek V4, GLM-5.1, Tab autocomplete, and more. ...

Qwen3.6 Series Now on SiliconFlow: Flagship-Level Coding in Compact Models

Built on community feedback, Qwen3.6 delivers substantial upgrades in coding agents, agentic workflows, and multimodal understanding with native thinking capabilities. ...

Qwen3.5 Series Now on SiliconFlow: Multimodal AI Excellence

The Qwen3.5 series, Alibaba's latest advancement in foundation model development, represents a significant leap forward in multimodal AI, featuring five powerful models ranging from 9B to 397B parameters. ...

Gemma 4 Now on SiliconFlow: Frontier Multimodal Intelligence

The Gemma 4 family of multimodal models by Google DeepMind delivers frontier-level performance, built for advanced reasoning and agentic workflows, and brings comprehensive multimodal capabilities and exceptional performance across diverse benchmarks. ...

DeepSeek-V4 Now on SiliconFlow: Million-Token Context Intelligence

DeepSeek-V4 introduces two powerful MoE models with groundbreaking 1M-token context windows, seting a new benchmark for unprecedented long-context efficiency, advanced reasoning capabilities, and SOTA performance across coding, mathematics, and agentic tasks. ...

Kimi K2.6 Now on SiliconFlow: SOTA Long-horizon Coding

Kimi K2.6, Moonshot AI's latest open-source multimodal agentic model. This release brings exceptional long-horizon coding capabilities, proactive autonomous execution, and swarm-based task orchestration, enabling developers to achieve production-grade agentic workflows with greater efficiency and reliability. ...

How to use Roo Code with SiliconFlow APIs

In this guide, we'll walk you through how to intergrate SiliconFlow APIs into Roo Code, giving you access to 200+ frontier models at competitive pricing. ...

How to use Cline with SiliconFlow APIs to Build Your rojects

In this guide, we'll show you how to integrate SiliconFlow's APIs into Cline, giving you access to frontier coding models like GLM-5.1, MiniMax M2.5, and Kimi K2.5 at cost-effective pricing. ...

How to use Chub AI with SiliconFlow's APIs

In this guide, we'll show you how to integrate SiliconFlow APIs into Chub AI, giving you access to 200+ top-tier models to power your RP experience. ...

GLM-5.1 Now on SiliconFlow: Autonomous Engineering at Scale

GLM-5.1, Z.AI's latest flagship model designed for long-horizon agentic engineering, achieves SOTA performance on SWE-Bench Pro and can work autonomously on complex tasks for up to 8 hours. ...

How to use Janitor AI with SiliconFlow APIs

In this guide, we'll show you how to use SiliconFlow APIs in Janitor AI, giving you access to 200+ open-source and frontier models at competitive pricing. ...

GLM-5V-Turbo Now on SiliconFlow: Vision Coding Redefined

GLM-5V-Turbo, Z. AI's multimodal coding foundation model built for vision-based coding and agent-driven tasks, brings native multimodal understanding and balanced visual-programming capabilities. ...

How to use Hermes Agent with SiliconFlow APIs

In this guide, we'll walk you through how to connect SiliconFlow's APIs to Hermes Agent, so you can power your agent with 200+ SOTA AI models in just a few steps. ...

MiniMax M2.5 Now on SiliconFlow: Built For Real-World Productivity

Discover MiniMax M2.5 — a frontier agentic model with SOTA coding, tool use, and office productivity, now accessible on SiliconFlow. ...

Step 3.5 Flash Now on SiliconFlow: Deep Thinking at Flash Speed

Step 3.5 Flash, the most capable StepFun AI's open-source foundation model is here. It is engineered to deliver frontier reasoning and agentic capabilities with exceptional efficiency. ...

GLM-5 Now on SiliconFlow: SOTA Open-Source Model Built for Agentic Engineering

Engineered for the era of agentic engineering, GLM-5 moves beyond 'vibe coding' to deliver reliable performance for complex systems and long-horizon tasks. ...

Kimi K2.5 Now on SiliconFlow: SOTA on Visual Agentic Intelligence

Built on Kimi K2 through continued pretraining over approximately 15T mixed visual and text tokens, it delivers state-of-the-art coding and vision capabilities as a native multimodal model. ...

MiniMax M2.1 on SiliconFlow: Enhanced Multi-Language Programming, Built for Real-World Complex Tasks

MiniMax M2.1, Open-source MoE model (10B active/230B total) built for real-world development—excels at multi-language programming, office automation, and agent workflows. ...

GLM-4.7 Now on SiliconFlow: Advanced Coding, Reasoning & Tool Use Capabilities

We're excited to announce that GLM-4.7, Z.ai's latest flagship model, is now available on SiliconFlow with Day 0 support. ...

FLUX.2 Now on SiliconFlow: Designed for Real-world Creative Workflows

We are excited to bring FLUX.2 [pro] and FLUX.2 [flex] to SiliconFlow, setting a new standard for real-world creative workflows. ...

GLM-4.6V Now on SiliconFlow: Native Multimodal Tool Use Meets SoTA Visual Intelligence

Z.ai's GLM-4.6V, featuring native function calling, 131K context, and SoTA visual understanding, is now available on SiliconFlow for building multimodal agents in real-world business scenarios. ...

Z-Image-Turbo Now on SiliconFlow: Photorealistic & Bilingual Text Rendering

Z-Image-Turbo, Alibaba Tongyi’s lightweight 6B text-to-image model, is now live on SiliconFlow, delivering photorealistic generation and bilingual text rendering on par with leading commercial models. ...

DeepSeek-V3.2 Now on SiliconFlow: Reasoning-first model built for agents

DeepSeek-V3.2 is now live on SiliconFlow, bringing GPT-5-level reasoning, efficient agentic performance, 164K context window and advanced tool-use to power your workflows. ...

Kimi K2 Thinking Now on SiliconFlow: Thinking Agent That Reasons and Acts

Designed as a thinking agent, it reasons step by step and can execute up to 200-300 sequential tool calls without human interference, thinking coherently across hundreds of steps to solve complex problems. ...

8 Key Insights on AI Infra from the co-founder of SiliconFlow

Pan Yang, co-founder of SiliconFlow, delivered a speech entitled “AI Infra: For Whom and Why?” at “Real-Time AI Infra Session” of Convo AI & RTE 2025. There are 8 core insights into the field of AI Infra. ...

MiniMax-M2 Now on SiliconFlow: Frontier-Style Coding and Agentic Intelligence

MiniMax-M2 is now live on SiliconFlow — a compact MoE model that delivers frontier-level reasoning and coding performance with efficiency and affordability. ...

Qwen3-VL-32B Now on SiliconFlow: Flagship-Level Intelligence with Dense-Model Efficiency

Qwen3-VL-32B is now live on SiliconFlow — a dense 32B model delivering flagship-level multimodal reasoning with faster responses, lower costs, and exceptional efficiency. ...

Qwen3-VL-8B Now on SiliconFlow: Small Model, Big Vision

Qwen3-VL-8B — the newest member of the Qwen3-VL family — is now live on SiliconFlow, offering full-scale multimodal reasoning at just 8B parameters. ...

Unlocking the Future of Video Technology: Introducing Hunyuan Video by Tencent

As video becomes central to communication and creativity, Tencent is reshaping the field with Hunyuan Video — an open-source AI platform built to revolutionize how we create, interact with, and share video content. ...

Zoom’s Transformation: Becoming an AI-First Company for the Future of Work

Work has evolved, and so has Zoom — now an AI-first company focused on smarter, faster collaboration beyond video calls, offering integrated tools that power the future of hybrid and remote work. ...

Ring-1T Now on SiliconFlow

Ring-1T Now on SiliconFlow: The World's First Open-Source Trillion-Parameter Thinking Model

Today, we're excited to bring Ring-1T to SiliconFlow — the first open-source trillion-parameter thinking model developed by Ant Group's inclusionAI team. ...

Ling-1T Now Live on SiliconFlow

Ling-1T Now Live on SiliconFlow: A Trillion-Scale Leap in Efficient Reasoning

Ling-1T — the first flagship model in the Ling 2.0 series — is now live on SiliconFlow, delivering state-of-the-art reasoning performance with an exceptional balance of accuracy, speed, and cost-efficiency. ...

Qwen3-VL on SiliconFlow

Qwen3-VL on SiliconFlow: Next-Gen VLM with Better World Understanding

Qwen3-VL, the most powerful vision-language model in the Qwen series, is now live on SiliconFlow — delivering advanced multimodal reasoning, 262K context, 32-language OCR, and up to 235B parameters for next-level AI understanding. ...

DeepSeek-V3.2-Exp Now on SiliconFlow

DeepSeek-V3.2-Exp Now on SiliconFlow: Efficient Long-Context Reasoning at 50% Lower Cost

DeepSeek-V3.2-Exp is now live on SiliconFlow — delivering faster, more efficient long-context reasoning and coding with DeepSeek Sparse Attention and a 164K context window. ...

Qwen3-Omni Now on SiliconFlow

Qwen3-Omni Now on SiliconFlow: Alibaba's Next-Gen Multimodal Foundation Model

Today, we're launching Qwen3-Omni on SiliconFlow — a native omni-modal foundation model that processes text, vision, audio, and video inputs with real-time streaming output, redefining multimodal AI capabilities. ...

GLM-4.6 Now on SiliconFlow: Advanced Agentic, Reasoning and Coding Capabilities

SiliconFlow is excited to bring you GLM-4.6, the latest version of Z.ai’s flagship model — featuring enhanced long-context reasoning, deeper agent integration, and improved coding efficiency. ...

Hunyuan-MT-7B Now On SiliconFlow

Hunyuan-MT-7B Now On SiliconFlow: High-quality Translation Model Supporting 33 Languages

SiliconFlow is thrilled to bring Hunyuan-MT-7B to our model catalog — Tencent's first open-source multilingual translation model supporting bidirectional translation across 33 major languages. ...

Ling-flash-2.0 Now on SiliconFlow

Ling-flash-2.0 Now on SiliconFlow: Flagship MoE Model Delivering SOTA Reasoning and High Efficiency

SiliconFlow is excited to bring you Ling-flash-2.0, the third MoE model under Ling 2.0 architecture, featuring a step forward in combining efficiency and reasoning ability. ...

Qwen-Image Is Here on SiliconFlow

Qwen-Image Is Here on SiliconFlow: Superior Text Rendering & Precise Image Editing

Today, Qwen-Image — a 20B MMDiT foundation model — is now available on SiliconFlow, offering advanced text rendering, precise editing, and versatile image generation. ...

Qwen-Image-Edit Now Available on SiliconFlow

Qwen-Image-Edit Now Available on SiliconFlow: Advanced Text, Semantic and Appearance Editing

Qwen-Image-Edit is now available on SiliconFlow. Powered by the Qwen-Image 20B model, it offers precise text editing and comprehensive semantic and appearance editing while preserving realism. ...

Ling-mini-2.0 Now on SiliconFlow

Ling-mini-2.0 Now on SiliconFlow: MoE Model with SOTA Performance & High Efficiency

SiliconFlow is excited to introduce Ling-mini-2.0 — Ant Group inclusionAI's MoE-based language model that redefines how efficient AI models can be. ...

Kimi K2-0905 Now on SiliconFlow

Kimi K2-0905 Now on SiliconFlow: Enhanced Coding Capabilities with Smoother Claude Code Integration

SiliconFlow is excited to welcome Kimi K2-0905 — Moonshot AI's most capable coding-focused upgrade that transforms how developers build and innovate. ...

Seed-OSS-36B-Instruct Now Available On SiliconFlow

Seed-OSS-36B-Instruct Now Available On SiliconFlow: Smarter AI That Thinks On-Demand

SiliconFlow is excited to bring Seed-OSS-36B-Instruct to our model catalog - ByteDance's revolutionary open-source model that puts AI reasoning control in your hands. ...

DeepSeek-V3.1 on SiliconFlow

DeepSeek-V3.1 on SiliconFlow: Hybrid Thinking, Smarter Tools and 164K Context Window

SiliconFlow is thrilled to bring DeepSeek-V3.1 to our model catalog — the latest upgrade from DeepSeek that takes AI one step closer to the Agent era. ...

penAI's gpt-oss Now Live on SiliconFlow

OpenAI's gpt-oss Now Live on SiliconFlow: Designed for Agentic Workflows, Advanced Reasoning and Tool Use

SiliconFlow is excited to announce the launch of gpt-oss-120B and gpt-oss-20B — state-of-the-art open-weight language models now available on our platform. ...

Wan 2.2 Series Now Available on SiliconFlow

Wan 2.2 Series Now Available on SiliconFlow：Generate Stable, Realistic and Cinematic Videos

SiliconFlow is excited to bring the Wan 2.2 series to our platform — a major upgrade to Wan's visual generative models, featuring significantly improved performance and superior visual quality. ...

GLM-4.5V: The World's Leading Open-Source Vision Reasoning Model Now on SiliconFlow

Today, we are excited to introduce that GLM-4.5V — the world’s best-performing open-source 100B-scale vision reasoning model — is now available on SiliconFlow. ...

Step3 Now Live on SiliconFlow: The Leading Open-source Multimodal Reasoning Model

Step3 Now Live on SiliconFlow: The Leading Open-source Multimodal Reasoning Model

Step3, Stepfun's latest cutting-edge multimodal reasoning model is now available on SiliconFlow. ...

Think Deeper, Act Faster: Qwen3-235B-A22B-Thinking-2507 Now Available on SiliconFlow

Think Deeper, Act Faster: Qwen3-235B-A22B-Thinking-2507 Now Available on SiliconFlow

With Qwen3-235B-A22B-Instruct-2507 (Non-Thinking mode) already demonstrating exceptional performance on SiliconFlow ...

GLM-4.5 Now Available on SiliconFlow

GLM-4.5 Now Available on SiliconFlow: Open-Source SOTA Model for Reasoning, Code, and Agentic Applications

Today, we're excited to integrate GLM-4.5 and GLM-4.5-Air, Z.ai's latest flagship model serie, into SiliconFlow platform. ...

Qwen3-235B-A22B-Instruct-2507: Leading Open-Source Model Now Live on SiliconFlow

Qwen3-235B-A22B-Instruct-2507: Leading Open-Source Model Now Live on SiliconFlow

Qwen has just released Qwen3-235B-A22B-Instruct-2507, an upgraded version of its flagship model Qwen3-235B-A22B Non-thinking. ...

Next-Gen Image Generation & Editing: FLUX.1 Kontext Now Available on SiliconFlow

Next-Gen Image Generation & Editing: FLUX.1 Kontext Now Available on SiliconFlow

SiliconFlow today introduces FLUX.1 Kontext [pro] and FLUX.1 Kontext [max], a suite of generative flow matching models that allows you to generate and edit images at the next level. ...

Kimi-K2 on SiliconFlow: Tailored for AI Agents, Priced to Scale

Kimi-K2 on SiliconFlow: Tailored for AI Agents, Priced to Scale

MoonShot AI's powerful open-source Kimi K2 Mixture-of-Experts (MoE) model is now available on SiliconFlow. ...

Run ERNIE-4.5-300B-A47B with SiliconFlow's API

The Baidu ERNIE team has announced the release of its ERNIE-4.5-300B-A47B, a powerful open-source large language model now available on the SiliconFlow platform. ...

Hunyuan-A13B-Instruct Now Available on SiliconFlow

The Tencent Hunyuan AI team announced the release of Hunyuan-A13B-Instruct, an open-source large language model (LLM) now available on the SiliconFlow platform. ...

FLUX.1 Kontext Dev on SiliconFlow：Redefining Ultra-Fast, Context-Aware Image Editing

FLUX.1 Kontext Dev on SiliconFlow：Redefining Ultra-Fast, Context-Aware Image Editing

FLUX.1 Kontext Dev, the world’s fastest and best context-aware open image editing model by Black Forest Labs, is now officially available on SiliconFlow. ...

MiniMax-M1-80k（456B）, the world’s first open-source hybrid-attention model at scale, is now available on SiliconFlow.

MiniMax-M1-80k Now Available on SiliconFlow

MiniMax-M1-80k（456B）, the world’s first open-source hybrid-attention model at scale, is now available on SiliconFlow. ...

DeepSeek-R1-0528 on SiliconFlow

Enhanced throughput, reduced hallucination, and improved reasoning — built for high-performance, high-reliability generative AI tasks. ...

Fast, Affordable, Powerful: Run Wan2.1 on SiliconFlow

Wan2.1, a comprehensive and open suite of video foundation models that pushes the boundaries of video generation, is now available on SiliconFlow. ...

Revolutionizing Spatial AI: Introducing World Labs’ 3D Generation Model

Generative AI has long focused on creating 2D content, but World Labs, co-founded by Fei-Fei Li, is taking a quantum leap forward — introducing a 3D generation model that turns a single image into an immersive, explorable world. ...

Unlocking Learning Potential: A Guide for Students on AI-Powered Academic Tools

Artificial intelligence is transforming education today—empowering students with personalized, efficient tools that enhance learning, boost productivity, and redefine how they achieve academic success. ...

DeepSeek-V3-0324 Is Now Live on SiliconFlow

DeepSeek-V3-0324 (671B) is now live on SiliconFlow — delivering major improvements in reasoning, writing, and math performance. ...

Mastering Creativity with Precision: How ControlNet Transforms Text-to-Image AI

AI-generated imagery is advancing rapidly, and ControlNet stands out as a breakthrough innovation — offering creators unprecedented precision and control in transforming text descriptions into highly customized visuals. ...

The AI Revolution in Advertising

The AI Revolution in Advertising: Transforming Marketing in the Digital Age

AI's rapid evolution is transforming advertising, as seen in Curious Refuge's first AI Advertising Competition, which showcased how AI-generated commercials are redefining creativity and marketing. ...

Introducing QwQ 32B-preview: A Breakthrough in Reasoning AI

In the fast-moving world of AI, Alibaba Cloud’s Qwen Team has introduced QwQ, an open-source experimental model pioneering human-like reasoning and redefining analytical intelligence. ...

Long Context vs. RAG: Which Should You Use for Large Documents?

Long context sends every document with the request. RAG sends the passages retrieval ranked highest. Compare token cost, updates, citations, and access control. ...

GLM-5.3-Flash vs. GLM-5.2 vs. Claude Opus 4.8: A Hands-On Coding and Agent Benchmark

GLM-5.3-Flash scores 63.4 on DeepSWE v1.1 against 46.2 for GLM-5.2, and lands within a point of Claude Opus 4.8 on Terminal-Bench. See the costs behind them. ...

GLM-5.3-Flash on SiliconFlow: Architecture, Pricing and Multimodal Limits

A 320B mixture-of-experts model with 18B active, 1M context, and text, image, video, and file input. Here is what one million tokens actually costs. ...

DeepSeek V4 Pro on SiliconFlow: Reasoning Levels, Pricing and Long-Task Cost

Low handles routine edits, High covers most multi-step work, and Max earns its cost on ambiguous failures. Route by task risk, not by habit. ...

What Is LLM Inference? Latency, Throughput, Batching, and Cost Explained

Prefill sets time to first token and decode sets the rest. See how TTFT, TPOT, batching, GPU memory, and serving choices change latency and cost. ...

What Is an AI Gateway? How It Manages Multiple LLM APIs

An AI gateway sits between your app and every model call. Compare it with a traditional API gateway, and see how routing, fallback, and quotas work. ...

RAG vs. Fine-Tuning: Which Should You Use for LLM Applications?

RAG adds outside knowledge at request time; fine-tuning changes how the model behaves. Compare data needs, freshness, latency, and cost before you pick. ...

Serverless Inference vs Reserved GPUs: Which Is Better for Your AI Workload?

Serverless inference bills per model call; reserved GPUs trade that flexibility for predictable capacity. Compare cost, scaling, and operational load. ...

How to Use an AI Video Generation API: Submit Jobs, Poll Status, and Store Results

Video generation outlasts a single HTTP request. See how requestId polling, bounded backoff, and object storage keep a long-running job reliable. ...

Do You Need a Reranker for RAG? Embeddings, Reranking, and Two-Stage Retrieval Explained

Most RAG failures are ranking problems, not retrieval gaps. Compare recall@k and nDCG, then decide whether a reranker earns its latency and cost. ...

SiliconFlow API Rate Limits and 429 Retry Strategies

SiliconFlow rate limits apply per account and per model, so bursts, shared keys or retries trigger 429 errors. Backoff, jitter and queues keep traffic stable. ...

OpenAI SDK Setup for SiliconFlow API Integration

Set the SiliconFlow base URL and model ID in the OpenAI SDK, then verify streaming, tools, vision, reasoning, limits, and errors before production. ...

Tencent Hy4 Preview: 1M Context, Coding, and API Guide

Evaluate Tencent Hy4 Preview’s 1M context, MoE architecture, coding benchmarks, deployment options, preview limits, and API cost factors in production. ...

DeepSeek V4.1 Flash: Architecture, Benchmarks, Pricing, and API Migration

Compare DeepSeek V4.1 Flash architecture, multimodal benchmarks, direct API pricing, migration steps, and cost controls for coding agents in production. ...

DeepSeek V4 Pro Function Calling: A Python Tutorial for Tool-Using Agents

DeepSeek V4 Pro Function Calling: A Python Tutorial for Tool-Using Agents

Build a DeepSeek V4 Pro tool-calling agent in Python. Validate function arguments, execute approved tools, return results, and handle errors safely. ...

How to Use DeepSeek V4 Pro in Claude Code: SiliconFlow Setup and Troubleshooting

How to Use DeepSeek V4 Pro in Claude Code: SiliconFlow Setup and Troubleshooting

Connect Claude Code to DeepSeek V4 Pro through SiliconFlow by setting the Anthropic base URL, mapping the model ID, and testing the Messages endpoint. ...

GLM-5.3 vs. GLM-5.3-Flash: Coding, Multimodal Input, and API Cost Compared

GLM-5.3 vs. GLM-5.3-Flash: Coding, Multimodal Input, and API Cost Compared

GLM-5.3 and GLM-5.3-Flash differ in coding strength, image input, and token pricing. Compare context limits and API cost before routing work between them. ...

DeepSeek V4 Pro vs. GLM 5.3: Which Should You Choose for Coding and Agents?

DeepSeek V4 Pro vs. GLM 5.3: Which Should You Choose for Coding and Agents?

DeepSeek V4 Pro costs less per token, while GLM 5.3 leads several long-horizon coding benchmarks. Compare pricing, context, and agent fit to choose. ...

GLM 5.3 API Guide: Pricing, Model ID, and Your First Request

GLM 5.3 API Guide: Pricing, Model ID, and Your First Request on SiliconFlow

GLM-5.3 on SiliconFlow costs $1.40 per 1M input tokens and $4.40 per 1M output tokens. See the model ID, a Python first request, streaming, and error fixes. ...

GLM-5.3 Now Live on SiliconFlow

A New Flagship Model for Complex Software Engineering and Long-Horizon Agent Tasks ...

Cover image for GLM-5.3 coding guide showing use cases, benchmarks, and limits for software engineering tasks

Is GLM-5.3 Good for Coding? Best Use Cases and Current Limits

GLM-5.3 improves terminal coding and long-horizon agent tasks. Review its strengths, benchmarks, practical limits, and SiliconFlow testing steps. ...

GLM-5.3 vs Claude Fable 5 coding benchmarks, API pricing, and context window comparison on SiliconFlow

GLM-5.3 vs Claude Fable 5 Coding Benchmarks and Cost

GLM-5.3 and Claude Fable 5 both target coding agents, but GLM-5.3 costs far less per token. Compare benchmarks, context windows, and API pricing to decide. ...

DeepSeek V4 Flash for Coding Agents — cover image showing AI coding assistant with code and lightning bolt

DeepSeek V4 Flash Coding Agent Use Cases and Pricing

DeepSeek V4 Flash excels at routine coding agent tasks. Route work based on validation results, task ambiguity, and cost per successful completion. ...

DeepSeek-V4-Pro-0813 Now Live on SiliconFlow

Enhanced Agent Capabilities for Production Workflows ...

FP8 Inference: What It Means for Your API Bill

FP8 can improve inference efficiency, but precision alone does not set API prices. Compare token rates, caching, retries, and cost per successful task. ...

Prompt Caching on SiliconFlow: How Cache Read Pricing Can Reduce Your API Bill

See how SiliconFlow cache read pricing changes API cost estimates, where prompt caching saves the most, and which workload factors affect real savings. ...

Best Cheap API for SillyTavern: DeepSeek-V4-Flash, GLM-5.2, and Long-Context Cost

Compare DeepSeek-V4-Flash and GLM-5.2 costs for long SillyTavern sessions, including context growth, token pricing, setup, and practical cost examples. ...

DeepSeek V4 Flash 0731 Now Live on SiliconFlow

The official release of DeepSeek V4 Flash, same size and architecture as the preview, re-post-trained for a significant jump in agentic capability. ...

Best API models for Cline compared across DeepSeek V4 Flash, DeepSeek V4 Pro, and GLM 5.2 for coding cost and performance

Models for Cline: DeepSeek V4 Flash, DeepSeek V4 Pro, and GLM 5.2 Cost Compared

Compare DeepSeek V4 Flash, DeepSeek V4 Pro, and GLM 5.2 as Cline API models, with cost examples and routing guidance for coding workflows. ...

Claude Opus alternative comparison showing GLM-5.2, DeepSeek-V4-Pro, DeepSeek-V4-Flash, Kimi-K2.6, and DeepSeek-V3.2 for coding workflows

Claude Opus Alternative for Coding: GLM-5.2 and DeepSeek-V4 Compared

Compare GLM-5.2, DeepSeek-V4-Pro, DeepSeek-V4-Flash, Kimi-K2.6, and DeepSeek-V3.2 as Claude Opus coding alternatives with pricing and context. ...

Claude Code alternatives comparison covering CLI agents, IDE assistants, open-source coding tools, and API-based developer workflows

Claude Code Alternatives: Open-Source, CLI, IDE, and API-Based Coding Tools Compared

Compare Claude Code alternatives across CLI, IDE, open-source, and API-based coding tools, and see how SiliconFlow powers flexible model access. ...

DeepSeek V4 API pricing comparison for Pro and Flash models with context window, cached input, and cost examples

DeepSeek V4 API Pricing: Pro vs Flash, Context Window, and Cost Examples

Compare DeepSeek V4 Pro and Flash API pricing, context window, and cached-input costs with worked examples for chatbots, RAG, and coding agents. ...

How to use Open Design with SiliconFlow's APIs

Use SiliconFlow as the built-in model provider in Open Design — the open-source, agent-native design workspace — to generate prototypes, landing pages, dashboards, decks, and more with 200+ models. ...

SiliconFlow-style cover image for Kimi model comparison guide

Kimi K3 vs. Kimi K2.7 Code vs. Kimi K2.6: Which Model Should You Choose?

Compare Kimi K3, Kimi K2.7 Code, and Kimi K2.6 by context length, coding fit, JSON support, multimodal use, latency, and task cost. ...

SiliconFlow-style cover image for Kimi K3 API on SiliconFlow

Kimi K3 on SiliconFlow: Features, Pricing, and API Access

Learn Kimi K3 features, SiliconFlow API access, pricing factors, context limits, JSON Mode, tool calling, and production planning checks. ...

Kimi K3 Now Live on SiliconFlow: The First Open 3T-Class Model at Frontier-Level Performance

Kimi K3 is now available on SiliconFlow — the first open 3T-class model with 2.8T parameters, 1M-token context, and native vision. OpenAI- and Anthropic-compatible API, starting at $3.0/M input tokens. Integrates with Claude Code, Cline, Cursor, and more. ...

Developer reviewing the Hy3 AI model dashboard with Mixture-of-Experts architecture, 256K context, and coding workflow tools

What Is Hy3? Tencent Hunyuan’s MoE Model Explained for Developers

Tencent's Hy3 MoE model leverages 295B total and 21B active parameters to power reasoning, coding, and agent workflows. See how its sparse architecture works. ...

Developer using a laptop and large monitor to build long-context AI agent workflows with code, document analysis, and API tools

Hy3 API Guide: How to Use Tencent Hunyuan Hy3 on SiliconFlow

Integrate the Tencent Hunyuan Hy3 API seamlessly for coding agents and long-context tasks. Claim your two weeks of free OpenAI-compatible access today. ...

Hy3 Now Live on SiliconFlow: Advancing Agent Capabilities, Ready for Real-World Workflows

Hy3, Tencent Hy Team's next-generation flagship, is now live on SiliconFlow. Free fro the first 2 weeks. ...

LongCat-2.0 Now on SiliconFlow: 1.6T MoE, Native 1M Context, Built for Agentic Coding

LongCat-2.0 is now live on SiliconFlow Meituan LongCat's next-generation trillion-parameter flagship model, and it's the full model behind "Owl Alpha", the anonymous model that reached Top 3 globally by daily volume on OpenRouter. ...

SiliconFlow Summer Rush丨GLM 5.2 Week : The More GLM 5.2 You Run, the More You Win

SiliconFlow Summer Rush — GLM 5.2 Week is Live ...

Team discussing GLM-5.2 vs GPT-5.5 in a modern office, with large screens showing AI model pricing, context window comparison, API access, and deployment options.

GLM-5.2 vs GPT-5.5: Pricing, Context Window & Open-Source Access

GLM-5.2 costs 77% to 87% less than GPT-5.5 for large workloads while offering an identical 1M token context window and MIT-licensed open weights. ...

Developer reviewing an AI model API dashboard with workflow, token flow, pricing analysis, and long context visualization on a large monitor in a modern office.

GLM 5.2 API Guide: GLM-5.2 Pricing, Model ID, 1M Context, and Cost Examples

Optimize your 1M context window costs. Learn how GLM-5.2 prompt caching reduces your SiliconFlow Serverless API bills with real examples. ...

Kimi K2.7 Code Now on SiliconFlow: Better Agentic Coding, 30% Fewer Thinking Tokens

Kimi K2.7 Code is Moonshot AI's latest open-source, coding-focused agentic model, now live on SiliconFlow. ...

GLM-5.2 Now on SiliconFlow: 1M Context, Long-Horizon Engineering, Near-Opus Coding

In this article, we have tested and compared GLM-5.2 with Opus 4.8, GPT-5.5, and GLM-5.1 under the same conditions, highlighting where GLM-5.2 reaches frontier-level performance while offering a major cost advantage on SiliconFlow. ...

Nex-N2-Pro Now on SiliconFlow: Agentic Thinking, Frontier Coding, and Long-Horizon Execution

Nex-N2-Pro is an agentic model with Agentic Thinking. Built for real-world productivity, it turns reasoning into executable, verifiable, and iterable action across agentic coding, deep research, tool calling, and terminal execution. ...

How to Use CodeWhale with SiliconFlow: DeepSeek V4 Terminal Coding Agent Setup

CodeWhale brings DeepSeek V4 to your terminal with a structured harness and recursive improvement that prevents model drift. With SiliconFlow as a native provider, setup takes a single command. ...

MiniMax M3 Now on SiliconFlow: Frontier Coding, 1M Context, and Native Multimodality

MiniMax M3, the first open-weight model to bring frontier coding, million-token context, and native multimodality together in one system, is now officially live on SiliconFlow. ...

How to Build a Personal Wiki Knowledge Base: Andrej Karpathy's Pattern with SiliconFlow & OpenCode

Build an automated wiki knowledge base based on Andrej Karpathy's llm-wiki sharing. Crawl, analyze, and compile with SiliconFlow, OpenCode, and oh-my-openagent. ...

How to Deploy an AI Assistant on Discord Using Hermes Agent

This guide takes you further, from choosing the right model on SiliconFlow to building your daily assistant on Discord. Plus, a set of practical tips and best practices to help you get more out of every session. ...

How to use CC Switch with SiliconFlow APIs

This guide will walk you through how to connect SiliconFlow to CC Switch and start powering your agentic workflows with 200+ frontier models in just a few steps. ...

How to use Continue in Vs Code with SiliconFlow APIs

Use SiliconFlow's frontier open-source models in Continue for VS Code. Step-by-step guide to configure DeepSeek V4, GLM-5.1, Tab autocomplete, and more. ...

Qwen3.6 Series Now on SiliconFlow: Flagship-Level Coding in Compact Models

Built on community feedback, Qwen3.6 delivers substantial upgrades in coding agents, agentic workflows, and multimodal understanding with native thinking capabilities. ...

Qwen3.5 Series Now on SiliconFlow: Multimodal AI Excellence

The Qwen3.5 series, Alibaba's latest advancement in foundation model development, represents a significant leap forward in multimodal AI, featuring five powerful models ranging from 9B to 397B parameters. ...

Gemma 4 Now on SiliconFlow: Frontier Multimodal Intelligence

The Gemma 4 family of multimodal models by Google DeepMind delivers frontier-level performance, built for advanced reasoning and agentic workflows, and brings comprehensive multimodal capabilities and exceptional performance across diverse benchmarks. ...

DeepSeek-V4 Now on SiliconFlow: Million-Token Context Intelligence

DeepSeek-V4 introduces two powerful MoE models with groundbreaking 1M-token context windows, seting a new benchmark for unprecedented long-context efficiency, advanced reasoning capabilities, and SOTA performance across coding, mathematics, and agentic tasks. ...

Kimi K2.6 Now on SiliconFlow: SOTA Long-horizon Coding

Kimi K2.6, Moonshot AI's latest open-source multimodal agentic model. This release brings exceptional long-horizon coding capabilities, proactive autonomous execution, and swarm-based task orchestration, enabling developers to achieve production-grade agentic workflows with greater efficiency and reliability. ...

How to use Roo Code with SiliconFlow APIs

In this guide, we'll walk you through how to intergrate SiliconFlow APIs into Roo Code, giving you access to 200+ frontier models at competitive pricing. ...

How to use Cline with SiliconFlow APIs to Build Your rojects

In this guide, we'll show you how to integrate SiliconFlow's APIs into Cline, giving you access to frontier coding models like GLM-5.1, MiniMax M2.5, and Kimi K2.5 at cost-effective pricing. ...

How to use Chub AI with SiliconFlow's APIs

In this guide, we'll show you how to integrate SiliconFlow APIs into Chub AI, giving you access to 200+ top-tier models to power your RP experience. ...

GLM-5.1 Now on SiliconFlow: Autonomous Engineering at Scale

GLM-5.1, Z.AI's latest flagship model designed for long-horizon agentic engineering, achieves SOTA performance on SWE-Bench Pro and can work autonomously on complex tasks for up to 8 hours. ...

How to use Janitor AI with SiliconFlow APIs

In this guide, we'll show you how to use SiliconFlow APIs in Janitor AI, giving you access to 200+ open-source and frontier models at competitive pricing. ...

GLM-5V-Turbo Now on SiliconFlow: Vision Coding Redefined

GLM-5V-Turbo, Z. AI's multimodal coding foundation model built for vision-based coding and agent-driven tasks, brings native multimodal understanding and balanced visual-programming capabilities. ...

How to use Hermes Agent with SiliconFlow APIs

In this guide, we'll walk you through how to connect SiliconFlow's APIs to Hermes Agent, so you can power your agent with 200+ SOTA AI models in just a few steps. ...

MiniMax M2.5 Now on SiliconFlow: Built For Real-World Productivity

Discover MiniMax M2.5 — a frontier agentic model with SOTA coding, tool use, and office productivity, now accessible on SiliconFlow. ...

Step 3.5 Flash Now on SiliconFlow: Deep Thinking at Flash Speed

Step 3.5 Flash, the most capable StepFun AI's open-source foundation model is here. It is engineered to deliver frontier reasoning and agentic capabilities with exceptional efficiency. ...

GLM-5 Now on SiliconFlow: SOTA Open-Source Model Built for Agentic Engineering

Engineered for the era of agentic engineering, GLM-5 moves beyond 'vibe coding' to deliver reliable performance for complex systems and long-horizon tasks. ...

Kimi K2.5 Now on SiliconFlow: SOTA on Visual Agentic Intelligence

Built on Kimi K2 through continued pretraining over approximately 15T mixed visual and text tokens, it delivers state-of-the-art coding and vision capabilities as a native multimodal model. ...

MiniMax M2.1 on SiliconFlow: Enhanced Multi-Language Programming, Built for Real-World Complex Tasks

MiniMax M2.1, Open-source MoE model (10B active/230B total) built for real-world development—excels at multi-language programming, office automation, and agent workflows. ...

GLM-4.7 Now on SiliconFlow: Advanced Coding, Reasoning & Tool Use Capabilities

We're excited to announce that GLM-4.7, Z.ai's latest flagship model, is now available on SiliconFlow with Day 0 support. ...

FLUX.2 Now on SiliconFlow: Designed for Real-world Creative Workflows

We are excited to bring FLUX.2 [pro] and FLUX.2 [flex] to SiliconFlow, setting a new standard for real-world creative workflows. ...

GLM-4.6V Now on SiliconFlow: Native Multimodal Tool Use Meets SoTA Visual Intelligence

Z.ai's GLM-4.6V, featuring native function calling, 131K context, and SoTA visual understanding, is now available on SiliconFlow for building multimodal agents in real-world business scenarios. ...

Z-Image-Turbo Now on SiliconFlow: Photorealistic & Bilingual Text Rendering

Z-Image-Turbo, Alibaba Tongyi’s lightweight 6B text-to-image model, is now live on SiliconFlow, delivering photorealistic generation and bilingual text rendering on par with leading commercial models. ...

DeepSeek-V3.2 Now on SiliconFlow: Reasoning-first model built for agents

DeepSeek-V3.2 is now live on SiliconFlow, bringing GPT-5-level reasoning, efficient agentic performance, 164K context window and advanced tool-use to power your workflows. ...

Kimi K2 Thinking Now on SiliconFlow: Thinking Agent That Reasons and Acts

Designed as a thinking agent, it reasons step by step and can execute up to 200-300 sequential tool calls without human interference, thinking coherently across hundreds of steps to solve complex problems. ...

8 Key Insights on AI Infra from the co-founder of SiliconFlow

Pan Yang, co-founder of SiliconFlow, delivered a speech entitled “AI Infra: For Whom and Why?” at “Real-Time AI Infra Session” of Convo AI & RTE 2025. There are 8 core insights into the field of AI Infra. ...

MiniMax-M2 Now on SiliconFlow: Frontier-Style Coding and Agentic Intelligence

MiniMax-M2 is now live on SiliconFlow — a compact MoE model that delivers frontier-level reasoning and coding performance with efficiency and affordability. ...

Qwen3-VL-32B Now on SiliconFlow: Flagship-Level Intelligence with Dense-Model Efficiency

Qwen3-VL-32B is now live on SiliconFlow — a dense 32B model delivering flagship-level multimodal reasoning with faster responses, lower costs, and exceptional efficiency. ...

Qwen3-VL-8B Now on SiliconFlow: Small Model, Big Vision

Qwen3-VL-8B — the newest member of the Qwen3-VL family — is now live on SiliconFlow, offering full-scale multimodal reasoning at just 8B parameters. ...

Unlocking the Future of Video Technology: Introducing Hunyuan Video by Tencent

As video becomes central to communication and creativity, Tencent is reshaping the field with Hunyuan Video — an open-source AI platform built to revolutionize how we create, interact with, and share video content. ...

Zoom’s Transformation: Becoming an AI-First Company for the Future of Work

Work has evolved, and so has Zoom — now an AI-first company focused on smarter, faster collaboration beyond video calls, offering integrated tools that power the future of hybrid and remote work. ...

Ring-1T Now on SiliconFlow

Ring-1T Now on SiliconFlow: The World's First Open-Source Trillion-Parameter Thinking Model

Today, we're excited to bring Ring-1T to SiliconFlow — the first open-source trillion-parameter thinking model developed by Ant Group's inclusionAI team. ...

Ling-1T Now Live on SiliconFlow

Ling-1T Now Live on SiliconFlow: A Trillion-Scale Leap in Efficient Reasoning

Ling-1T — the first flagship model in the Ling 2.0 series — is now live on SiliconFlow, delivering state-of-the-art reasoning performance with an exceptional balance of accuracy, speed, and cost-efficiency. ...

Qwen3-VL on SiliconFlow

Qwen3-VL on SiliconFlow: Next-Gen VLM with Better World Understanding

Qwen3-VL, the most powerful vision-language model in the Qwen series, is now live on SiliconFlow — delivering advanced multimodal reasoning, 262K context, 32-language OCR, and up to 235B parameters for next-level AI understanding. ...

DeepSeek-V3.2-Exp Now on SiliconFlow

DeepSeek-V3.2-Exp Now on SiliconFlow: Efficient Long-Context Reasoning at 50% Lower Cost

DeepSeek-V3.2-Exp is now live on SiliconFlow — delivering faster, more efficient long-context reasoning and coding with DeepSeek Sparse Attention and a 164K context window. ...

Qwen3-Omni Now on SiliconFlow

Qwen3-Omni Now on SiliconFlow: Alibaba's Next-Gen Multimodal Foundation Model

Today, we're launching Qwen3-Omni on SiliconFlow — a native omni-modal foundation model that processes text, vision, audio, and video inputs with real-time streaming output, redefining multimodal AI capabilities. ...

GLM-4.6 Now on SiliconFlow: Advanced Agentic, Reasoning and Coding Capabilities

SiliconFlow is excited to bring you GLM-4.6, the latest version of Z.ai’s flagship model — featuring enhanced long-context reasoning, deeper agent integration, and improved coding efficiency. ...

Hunyuan-MT-7B Now On SiliconFlow

Hunyuan-MT-7B Now On SiliconFlow: High-quality Translation Model Supporting 33 Languages

SiliconFlow is thrilled to bring Hunyuan-MT-7B to our model catalog — Tencent's first open-source multilingual translation model supporting bidirectional translation across 33 major languages. ...

Ling-flash-2.0 Now on SiliconFlow

Ling-flash-2.0 Now on SiliconFlow: Flagship MoE Model Delivering SOTA Reasoning and High Efficiency

SiliconFlow is excited to bring you Ling-flash-2.0, the third MoE model under Ling 2.0 architecture, featuring a step forward in combining efficiency and reasoning ability. ...

Qwen-Image Is Here on SiliconFlow

Qwen-Image Is Here on SiliconFlow: Superior Text Rendering & Precise Image Editing

Today, Qwen-Image — a 20B MMDiT foundation model — is now available on SiliconFlow, offering advanced text rendering, precise editing, and versatile image generation. ...

Qwen-Image-Edit Now Available on SiliconFlow

Qwen-Image-Edit Now Available on SiliconFlow: Advanced Text, Semantic and Appearance Editing

Qwen-Image-Edit is now available on SiliconFlow. Powered by the Qwen-Image 20B model, it offers precise text editing and comprehensive semantic and appearance editing while preserving realism. ...

Ling-mini-2.0 Now on SiliconFlow

Ling-mini-2.0 Now on SiliconFlow: MoE Model with SOTA Performance & High Efficiency

SiliconFlow is excited to introduce Ling-mini-2.0 — Ant Group inclusionAI's MoE-based language model that redefines how efficient AI models can be. ...

Kimi K2-0905 Now on SiliconFlow

Kimi K2-0905 Now on SiliconFlow: Enhanced Coding Capabilities with Smoother Claude Code Integration

SiliconFlow is excited to welcome Kimi K2-0905 — Moonshot AI's most capable coding-focused upgrade that transforms how developers build and innovate. ...

Seed-OSS-36B-Instruct Now Available On SiliconFlow

Seed-OSS-36B-Instruct Now Available On SiliconFlow: Smarter AI That Thinks On-Demand

SiliconFlow is excited to bring Seed-OSS-36B-Instruct to our model catalog - ByteDance's revolutionary open-source model that puts AI reasoning control in your hands. ...

DeepSeek-V3.1 on SiliconFlow

DeepSeek-V3.1 on SiliconFlow: Hybrid Thinking, Smarter Tools and 164K Context Window

SiliconFlow is thrilled to bring DeepSeek-V3.1 to our model catalog — the latest upgrade from DeepSeek that takes AI one step closer to the Agent era. ...

penAI's gpt-oss Now Live on SiliconFlow

OpenAI's gpt-oss Now Live on SiliconFlow: Designed for Agentic Workflows, Advanced Reasoning and Tool Use

SiliconFlow is excited to announce the launch of gpt-oss-120B and gpt-oss-20B — state-of-the-art open-weight language models now available on our platform. ...

Wan 2.2 Series Now Available on SiliconFlow

Wan 2.2 Series Now Available on SiliconFlow：Generate Stable, Realistic and Cinematic Videos

SiliconFlow is excited to bring the Wan 2.2 series to our platform — a major upgrade to Wan's visual generative models, featuring significantly improved performance and superior visual quality. ...

GLM-4.5V: The World's Leading Open-Source Vision Reasoning Model Now on SiliconFlow

Today, we are excited to introduce that GLM-4.5V — the world’s best-performing open-source 100B-scale vision reasoning model — is now available on SiliconFlow. ...

Step3 Now Live on SiliconFlow: The Leading Open-source Multimodal Reasoning Model

Step3 Now Live on SiliconFlow: The Leading Open-source Multimodal Reasoning Model

Step3, Stepfun's latest cutting-edge multimodal reasoning model is now available on SiliconFlow. ...

Think Deeper, Act Faster: Qwen3-235B-A22B-Thinking-2507 Now Available on SiliconFlow

Think Deeper, Act Faster: Qwen3-235B-A22B-Thinking-2507 Now Available on SiliconFlow

With Qwen3-235B-A22B-Instruct-2507 (Non-Thinking mode) already demonstrating exceptional performance on SiliconFlow ...

GLM-4.5 Now Available on SiliconFlow

GLM-4.5 Now Available on SiliconFlow: Open-Source SOTA Model for Reasoning, Code, and Agentic Applications

Today, we're excited to integrate GLM-4.5 and GLM-4.5-Air, Z.ai's latest flagship model serie, into SiliconFlow platform. ...

Qwen3-235B-A22B-Instruct-2507: Leading Open-Source Model Now Live on SiliconFlow

Qwen3-235B-A22B-Instruct-2507: Leading Open-Source Model Now Live on SiliconFlow

Qwen has just released Qwen3-235B-A22B-Instruct-2507, an upgraded version of its flagship model Qwen3-235B-A22B Non-thinking. ...

Next-Gen Image Generation & Editing: FLUX.1 Kontext Now Available on SiliconFlow

Next-Gen Image Generation & Editing: FLUX.1 Kontext Now Available on SiliconFlow

SiliconFlow today introduces FLUX.1 Kontext [pro] and FLUX.1 Kontext [max], a suite of generative flow matching models that allows you to generate and edit images at the next level. ...

Kimi-K2 on SiliconFlow: Tailored for AI Agents, Priced to Scale

Kimi-K2 on SiliconFlow: Tailored for AI Agents, Priced to Scale

MoonShot AI's powerful open-source Kimi K2 Mixture-of-Experts (MoE) model is now available on SiliconFlow. ...

Run ERNIE-4.5-300B-A47B with SiliconFlow's API

The Baidu ERNIE team has announced the release of its ERNIE-4.5-300B-A47B, a powerful open-source large language model now available on the SiliconFlow platform. ...

Hunyuan-A13B-Instruct Now Available on SiliconFlow

The Tencent Hunyuan AI team announced the release of Hunyuan-A13B-Instruct, an open-source large language model (LLM) now available on the SiliconFlow platform. ...

FLUX.1 Kontext Dev on SiliconFlow：Redefining Ultra-Fast, Context-Aware Image Editing

FLUX.1 Kontext Dev on SiliconFlow：Redefining Ultra-Fast, Context-Aware Image Editing

FLUX.1 Kontext Dev, the world’s fastest and best context-aware open image editing model by Black Forest Labs, is now officially available on SiliconFlow. ...

MiniMax-M1-80k（456B）, the world’s first open-source hybrid-attention model at scale, is now available on SiliconFlow.

MiniMax-M1-80k Now Available on SiliconFlow

MiniMax-M1-80k（456B）, the world’s first open-source hybrid-attention model at scale, is now available on SiliconFlow. ...

DeepSeek-R1-0528 on SiliconFlow

Enhanced throughput, reduced hallucination, and improved reasoning — built for high-performance, high-reliability generative AI tasks. ...

Fast, Affordable, Powerful: Run Wan2.1 on SiliconFlow

Wan2.1, a comprehensive and open suite of video foundation models that pushes the boundaries of video generation, is now available on SiliconFlow. ...

Revolutionizing Spatial AI: Introducing World Labs’ 3D Generation Model

Generative AI has long focused on creating 2D content, but World Labs, co-founded by Fei-Fei Li, is taking a quantum leap forward — introducing a 3D generation model that turns a single image into an immersive, explorable world. ...

Unlocking Learning Potential: A Guide for Students on AI-Powered Academic Tools

Artificial intelligence is transforming education today—empowering students with personalized, efficient tools that enhance learning, boost productivity, and redefine how they achieve academic success. ...

DeepSeek-V3-0324 Is Now Live on SiliconFlow

DeepSeek-V3-0324 (671B) is now live on SiliconFlow — delivering major improvements in reasoning, writing, and math performance. ...

Mastering Creativity with Precision: How ControlNet Transforms Text-to-Image AI

AI-generated imagery is advancing rapidly, and ControlNet stands out as a breakthrough innovation — offering creators unprecedented precision and control in transforming text descriptions into highly customized visuals. ...

The AI Revolution in Advertising

The AI Revolution in Advertising: Transforming Marketing in the Digital Age

AI's rapid evolution is transforming advertising, as seen in Curious Refuge's first AI Advertising Competition, which showcased how AI-generated commercials are redefining creativity and marketing. ...

Introducing QwQ 32B-preview: A Breakthrough in Reasoning AI

In the fast-moving world of AI, Alibaba Cloud’s Qwen Team has introduced QwQ, an open-source experimental model pioneering human-like reasoning and redefining analytical intelligence. ...

Ready to accelerate your AI development?

Ready to accelerate your AI development?

Ready to accelerate your AI development?

PAGES

Home

Models

Products

Docs

Pricing

Blog

Articles

About

Contact

MODELS

Image

Video

LLM

Audio

PRODUCTS

Overview

Inference

Fine-tuning

Reserved GPUs

© 2026 SiliconFlow

Terms

·

Privacy

PAGES

Home

Models

Products

Docs

Pricing

Blog

Articles

About

Contact

MODELS

Image

Video

LLM

Audio

PRODUCTS

Overview

Inference

Fine-tuning

Reserved GPUs

© 2026 SiliconFlow

Terms

·

Privacy

PAGES

Home

Models

Products

Docs

Pricing

Blog

Articles

About

Contact

MODELS

Image

Video

LLM

Audio

PRODUCTS

Overview

Inference

Fine-tuning

Reserved GPUs

© 2026 SiliconFlow

Terms

·

Privacy