DeepSeek has released an updated version of its open-source large language model, DeepSeek V4.1 Flash 1017, on October 17, 2026. This iteration features a significant improvement across various benchmarks, particularly in coding and agent-based tasks. The model now boasts a 1M token context window and is available under the MIT License, allowing for commercial use. Key advancements include enhanced performance on coding benchmarks like DeepSWE v1.1 and Terminal-Bench 4.0, surpassing competitors such as Anthropic's Opus 5 and OpenAI's GPT-5.6 Sol. AI
IMPACT Sets new SOTA on coding benchmarks, potentially accelerating enterprise adoption of advanced agent frameworks.
RANK_REASON Frontier-lab model release with system card. [lever_c_demoted from frontier_release: ic=1 ai=1.0]
- Agents Last Exam
- DeepSeek
- DeepSWE v1.1
- GLM 5.3
- GPT-5.6 "Sol"
- Hugging Face
- Kimi k3
- MIT License
- NL2Repo-Bench
- ProgramBench
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →