DeepSeek releases open-weight V4.1-Flash-1017, beats GPT-5.6 Sol on agent benchmarks
victormustar · x · 2026-10-17
DeepSeek has open-sourced DeepSeek-V4.1-Flash-1017 on Hugging Face under an MIT license, with image-text-to-text support and fp8/8-bit weights.
The accompanying scorecard shows gains across the board over the previous V4.1-Flash and several wins over GPT-5.6 Sol:
- Terminal-Bench 4.0: 42.3 (vs 31.2 previously)
- Automation-Bench v1.0.6: 60.5 (vs 54.8)
- Apex Bench v1.1: 72.1 (vs 64.6)
- Agents' Last Exam: 35.4, above GLM-5.3, GPT-5.6 Sol and Opus 5
- ProgramBench: 29.1 vs GPT-5.6 Sol's 23.0
- ZeroBench-main hits 51.0, still below GPT-5.6 Sol (53.0) and Opus 5 (52.0)
More from Models
- Codex upgrades DSH and live-tests DeepSeek-V4.1-Flash 1017 — cedric_chee · 2026-10-17
- Will Human Brains Soon Fail to Judge Whether New Models Are Actually Better? — arjunrajlab · 2026-10-17
- User Reports Astra Agent Unprompted Finds Sophisticated Security Bugs via Gmail — Tasty-Trip5518 · 2026-10-17
- Qwen3.6 35B MoE hits 19.5 tok/s on a 16GB 4060 Ti, 10x faster than dense models — dreamyrhodes · 2026-10-17
- Official vs open benchmarks disagree wildly — a call for an open calibration benchmark — lahwran_ · 2026-10-17
- DeepSeek flash 1017 called out for still acting "very much like Claude" — teortaxesTex · 2026-10-17