DeepSeek
Text Generation
DeepSeek-V3.2
DeepSeek-V3.2 是一個模型,能夠將高計算效率與卓越的推理和代理性能相結合。它的方法建立在三個關鍵技術突破之上:DeepSeek Sparse Attention (DSA),這是一種有效的注意力機制,顯著降低了計算複雜性,同時保持模型性能,特別針對長上下文場景進行了優化;一個可擴展的強化學習框架,使其性能可與 GPT-5 比肩,推理能力則可與其高計算版本的 Gemini-3.0-Pro 並駕齊驅;以及一個大規模代理任務合成管道,用於在使用工具的場景中整合推理,提高在複雜交互環境中的合規性和泛化能力。該模型在 2025 年國際數學奧林匹克(IMO)和國際信息學奧林匹克(IOI)中獲得金牌成績。...
總上下文:
164K
最大輸出:
164K
輸入:
$
0.27
/ M Tokens
輸入:
$
text
/ M Tokens
輸出:
$
0.42
/ M Tokens
DeepSeek
Text Generation
DeepSeek-V3.2-Exp
DeepSeek-V3.2-Exp 是一個實驗版本的 DeepSeek 模型,基於 V3.1-Terminus 構建。它推出 DeepSeek Sparse Attention (DSA),能夠在長上下文上進行更快、更有效的訓練和推理。...
總上下文:
164K
最大輸出:
164K
輸入:
$
0.27
/ M Tokens
輸入:
$
text
/ M Tokens
輸出:
$
0.41
/ M Tokens
DeepSeek
Text Generation
DeepSeek-V3.1-Terminus
DeepSeek-V3.1-Terminus 是一個更新版本,基於 V3.1 的優勢,同時回應了用戶的關鍵反饋。它改善了語言的一致性,減少了中英文字混雜的情況以及偶爾出現的異常字符。此外,在代碼代理和搜尋代理的性能上也進行了升級。...
總上下文:
164K
最大輸出:
164K
輸入:
$
0.27
/ M Tokens
輸入:
$
text
/ M Tokens
輸出:
$
1.0
/ M Tokens
DeepSeek
Text Generation
DeepSeek-V3.1
DeepSeek-V3.1 是一個混合模型,支持思考模式和非思考模式。通過後期訓練優化,模型在工具使用和代理任務中的性能顯著提高。DeepSeek-V3.1-Think 的答案質量達到與 DeepSeek-R1-0528 相當的水平,同時回應速度更快。...
總上下文:
164K
最大輸出:
164K
輸入:
$
0.27
/ M Tokens
輸入:
$
text
/ M Tokens
輸出:
$
1.0
/ M Tokens
DeepSeek
Text Generation
DeepSeek-V3
DeepSeek-V3-0324 在幾個關鍵方面展示了相對其前代產品 DeepSeek-V3 的顯著改進,包括推理性能的重大提升、更強的前端開發能力以及更智能的工具使用能力。...
總上下文:
164K
最大輸出:
164K
輸入:
$
0.25
/ M Tokens
輸入:
$
text
/ M Tokens
輸出:
$
1.0
/ M Tokens
DeepSeek
Text Generation
DeepSeek-R1
DeepSeek-R1-0528 是一個升級的模型,顯示出在處理複雜推理任務方面的顯著改進,也提供了降低的幻覺率、增強的函數調用支持以及更好的振動編碼體驗。其性能可與 O3 和 Gemini 2.5 Pro 相媲美。...
總上下文:
164K
最大輸出:
164K
輸入:
$
0.5
/ M Tokens
輸入:
$
text
/ M Tokens
輸出:
$
2.18
/ M Tokens
DeepSeek
Text Generation
DeepSeek-V4.1-Flash
DeepSeek-V4.1-Flash is DeepSeek's latest multimodal MoE model with 552B backbone parameters. Built on a Causal Encoder-Decoder architecture, it activates only ~8B parameters during prefill and ~16B during decode, delivering flagship-level intelligence at significantly lower inference cost. The model natively understands images, supports a 1M-token context window. It outperforms DeepSeek-V4-Pro across coding, agentic and automation benchmarks, while reducing the KV cache footprint by roughly 4x versus its predecessor. ...
總上下文:
1049K
最大輸出:
393K
輸入:
$
0.15
/ M Tokens
輸入:
$
text
/ M Tokens
輸出:
$
0.6
/ M Tokens
DeepSeek
Text Generation
DeepSeek-V4-Flash-Vision-Exp
DeepSeek-V4-Flash-Vision-Exp is the first experimental multimodal model in the DeepSeek-V4 family. It builds on the DeepSeek-V4-Flash architecture by incorporating visual modules and undergoing continued training to unlock visual understanding capabilities. It matches the official DeepSeek-V4-Flash on text-only capabilities (agents, reasoning, world knowledge), while achieving substantial improvements on multimodal agent benchmarks, bringing performance close to Opus-4.8. It shares the same model structure as DeepSeek-V4-Flash-DSpark, i.e. it comes with a speculative decoding module attached. Images are tokenized for billing at up to 384 tokens each....
總上下文:
1049K
最大輸出:
393K
輸入:
$
0.44
/ M Tokens
輸入:
$
text
/ M Tokens
輸出:
$
1.32
/ M Tokens
DeepSeek
Text Generation
DeepSeek-V4-Pro-0813
DeepSeek-V4-Pro-0813 is the official release of DeepSeek-V4-Pro, superseding the preview version, with greatly enhanced agentic capabilities and performance improvements that are especially pronounced in production environments. It is built on the DeepSeek-V4-Pro (Preview) model structure, with a DSpark speculative decoding module attached....
總上下文:
1049K
最大輸出:
393K
輸入:
$
1.32
/ M Tokens
輸入:
$
text
/ M Tokens
輸出:
$
3.96
/ M Tokens
DeepSeek
Text Generation
DeepSeek-V4-Flash-0731
DeepSeek-V4-Flash-0731 is the official release of DeepSeek-V4-Flash, superseding the preview version, with substantially enhanced agentic capabilities. It has the same model structure as DeepSeek-V4-Flash-DSpark, i.e. it comes with a speculative decoding module attached....
總上下文:
1049K
最大輸出:
393K
輸入:
$
0.22
/ M Tokens
輸入:
$
text
/ M Tokens
輸出:
$
0.66
/ M Tokens
DeepSeek
Text Generation
DeepSeek-V4-Pro
DeepSeek-V4-Pro is DeepSeek's flagship open-source MoE model with 1.6T total parameters and 49B activated, purpose-built for frontier-level reasoning, coding, and agentic tasks. Supporting a 1M-token context window and three reasoning effort modes up to Think Max, it achieves top-tier performance on coding benchmarks such as LiveCodeBench and Codeforces — rivaling leading closed-source models — and is released under the MIT License....
總上下文:
1049K
最大輸出:
393K
輸入:
$
1.50162
/ M Tokens
輸入:
$
text
/ M Tokens
輸出:
$
3.135
/ M Tokens
DeepSeek
Text Generation
DeepSeek-V4-Flash
DeepSeek-V4-Flash is DeepSeek's latest open-source MoE model featuring 284B total parameters with only 13B activated during inference, delivering high-speed generation without sacrificing capability. With native support for a 1M-token context window and three switchable reasoning modes — Non-Think, Think High, and Think Max — it offers flexible intelligence scaling from everyday tasks to complex reasoning, all under the MIT License....
總上下文:
1049K
最大輸出:
393K
輸入:
$
0.13
/ M Tokens
輸入:
$
text
/ M Tokens
輸出:
$
0.28
/ M Tokens
