DeepSeek has released an updated version of its V4.1-Flash model, named DeepSeek V4.1-Flash-1017. This new open-source model, available under the MIT license on Hugging Face, claims significant improvements in coding benchmarks and a substantial reduction in cost per task compared to its predecessor. While DeepSeek's reported performance metrics show it outperforming models like Kimi K3 and GPT-5.6 Sol on several benchmarks, it still trails Claude Opus-5 on some tests. The cost-effectiveness of the model is verifiable through its public API pricing, but the performance claims await independent reproduction. AI
IMPACT Sets new SOTA on coding benchmarks; pressures Anthropic to respond.
RANK_REASON Frontier-lab model release with system card
Read on Mastodon — fosstodon.org →
- Artificial Analysis
- Claude Opus-5
- DeepSeek
- FalkorDB
- GPT-5.6 Sol
- Hugging Face
- Javascript
- Kimi K3
- Lakshmi
- MIT License
- ProgramBench
AI-generated summary · Google Gemini · from 6 sources. How we write summaries →