Nex-N2-Pro

Nex-N2-Pro

關於Nex-N2-Pro

Nex-N2 is a family of thinking models with Agentic Thinking. They adaptively decide when and how deeply to reason, unifying agent cognition across coding, search, and tool use into a single coherent paradigm. Key Claims - SOTA among open models on SWE-Verified, SWE-Pro, Terminal Bench 2.0, Tau3, WildClawBench, BFCL V4 - Top-tier in agentic coding (end-to-end dev loops), deep search (BrowserComp, Wild Search, FinSearch), and real-world productivity (GDP Val) - Adaptive Thinking: auto-adjusts reasoning depth per step, 30-50% fewer thinking tokens vs always-on, with equal or better performance - Plug-and-play with Claude Code, Cursor, OpenClaw, and agentic harnesses

Harness Nex-N2-Pro’s Agentic Thinking to automate complex workflows, from autonomous coding to deep market research.

Autonomous Software Engineering

Execute end-to-end development cycles, including debugging, testing, and deployment across complex codebases.

Use Case Example:

"Fixed a memory leak in a Rust-based data pipeline by tracing ownership issues and implementing a more efficient buffer management strategy."

Deep Research & Intelligence

Perform multi-step web searches and data synthesis to generate high-fidelity reports on technical or financial topics.

Use Case Example:

"Conducted a deep-dive into global semiconductor supply chains, synthesizing data from trade journals and earnings calls to predict 2026 price volatility."

Intelligent Tool Orchestration

Seamlessly integrate with APIs and external tools to automate multi-stage business processes and administrative tasks.

Use Case Example:

"Managed a cross-platform marketing campaign by coordinating between Trello, Slack, and Google Analytics to adjust ad spend based on real-time conversion data."

Adaptive DevOps & SRE

Navigate terminal environments and cloud infrastructure to diagnose issues and optimize system performance autonomously.

Use Case Example:

"Resolved a database latency spike in a PostgreSQL cluster by analyzing query plans and reconfiguring indexing strategies via the CLI."

元數據

創建於

許可證

APACHE-2.0

供應商

Nex AGI

HuggingFace

規格

狀態

Deprecated

架構

Transformer MoE

經過校準的

否

專家並行

是

總參數

397B

啟用的參數

推理

否

精度

FP8

上下文長度

262K

最大輸出長度

準備好 加速您的人工智能開發了嗎?

準備好 加速您的人工智能開發了嗎?

準備好 加速您的人工智能開發了嗎?

© 2025 SiliconFlow

© 2025 SiliconFlow

© 2025 SiliconFlow