DeepSeek-V4.1-Flash Base places in the top three on BigCodeBench.
DeepSeek Released September 10th, 2026 Family DeepSeek V4.1 The lab's model card Compare with others DeepSeek Sep 2026 Open weights Flagship
Its best score on each benchmark, and where that ranks among every model with a score there. We rank each model once, at its best setting.
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| AGIEval | 83.4% | #5 of 40 | Lab model card | |
| C-Eval | 92.1% | #6 of 46 | Lab model card | |
| DROP | 87.9% | #12 of 41 | Lab model card | |
| BIG-Bench Hard | 86.1% | #24 of 73 | Lab model card | |
| HellaSwag | 87.2% | #19 of 53 | Lab model card | |
| SimpleQA Verified | 42.3% | #12 of 22 | Lab model card | |
| MGSM | 80.2% | #35 of 57 | Lab model card | |
| BBEH micro average | 27.2% | #10 of 15 | Lab model card | |
| MMLU-Pro | 74.1% | #113 of 167 | Lab model card | |
| SuperGPQA | 53.1% | #27 of 35 | Lab model card | |
| LongBench v2 | 45.2% | #30 of 35 | Lab model card | |
| MultiLoKo | 45.5% | #3 of 3 | Lab model card |
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| BigCodeBench | 60.6% | #2 of 14 | Lab model card | |
| HumanEval | 79.4% | #56 of 101 | Lab model card |
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| DocVQA | 95.6% | #7 of 36 | Lab model card | |
| MMMU-Pro | 56.5% | #73 of 90 | Lab model card | |
| RefCOCO / RefCOCO+ / RefCOCOg | 86% | #24 of 29 | Lab model card | |
| CVBench | 77.9% | #14 of 15 | Lab model card |
The Self-Reported Index fits one score to the 19 capability results DeepSeek reported for this model, on the scale of the Epoch Capabilities Index. Labs choose which benchmarks to show and how to run them, so read it as the lab's claim, not an independent measurement. The 90% interval runs from 126.0 to 132.7. How we compute it.
Every result in the longer documents the lab published with it. 20 of them aren't in the launch post.