DeepSeekDeepSeek-V4.1
DeepSeek V4.1 Flash
DeepSeek · DeepSeek-V4.1. Released Sep 10, 2026.
DeepSeek V4.1 Flash is a model from DeepSeek in the DeepSeek-V4.1 family, released Sep 10, 2026. evals.report tracks 3 reported DeepSeek V4.1 Flash benchmark scores across AA-Omniscience: Knowledge and Hallucination Benchmark, Terminal-Bench 2.1, Humanity's Last Exam — each shown with its benchmark, metric, source status, and date, and never combined into a single ranking.
Open3 results
Benchmark results 3
Compare this model| Benchmark | Category | Score | Metric | Status | Date | |
|---|---|---|---|---|---|---|
| AA-Omniscience: Knowledge and Hallucination Benchmark | Reasoning | -5 | AA-Omniscience Index | Official | Sep 10, 2026 | Details |
| Terminal-Bench 2.1 | Agents | 90.6% | task success | Verified | Sep 10, 2026 | Details |
| Humanity's Last Exam | Reasoning | 36.8% | accuracy | Verified | Sep 10, 2026 | Details |