DeepSeek's 'minor update' jumps Terminal-Bench 4.0 from 31.2 to 42.3, MIT-licensed
mishig25 · x · 2026-10-17
DeepSeek released DeepSeek-V4.1-Flash-1017, calling it a Pareto improvement in performance and cost: stronger benchmarks, faster task completion, and shorter outputs. Live on web, app and API.
Key numbers from the quoted commentary:
- Terminal-Bench 4.0 jumped from 31.2 to 42.3 — the commenter notes it's funny how DeepSeek keeps calling these jumps "minor updates"
- Reports 77.0 on DeepSWE, vs 74.0 for Opus 5 and 73.0 for GPT-5.6 Sol
- The most interesting line: faster task completion with shorter outputs — using fewer tokens is a practical way the cost curve collapses, often treated as an afterthought
- Weights are MIT-licensed, in DeepSeek tradition
Caveat: it remains to be seen whether these numbers hold up outside DeepSeek's own eval setup.
More from Models
- Anthropic says Claude-designed quantum circuits cut cost of key operation ~17x — Dr_Singularity · 2026-10-17
- Training-free Confident Decoding skips final layers, cuts alignment tax, boosts reasoning — aiamblichus · 2026-10-17
- Gemini Pro subscriber details crash bugs and forced model downgrades in app — Kindly_Rough_4300 · 2026-10-17
- 'No, the whale didn't distill Sol': DeepSeek-V4.1-Flash comparison sparks memes — cedric_chee · 2026-10-17
- Gemini 3 vs Opus 4.5 vs GPT-5.1 vs Grok 4: same-prompt 3D diorama showdown — cedric_chee · 2026-10-17
- DeepSeek's monthly release cadence pressures Zai and Moonshot's premium pricing — zephyr_z9 · 2026-10-17