DeepSeek V4.1 Flash (1017): benchmark results
DeepSeek's open-weight V4.1 Flash 1017 (October 17, 2026), an MIT-licensed update to V4.1 Flash on the same 552B-parameter architecture. Provider: DeepSeek. Released 2026-10-17. Access: Open.
Aggregate Rating 1911 ± 21, rank #15 of 1648 rated models, from 42 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| DeepSeek V4.1 Flash 1017 Card - Agents' Last Exam | 35.4 | Score (%) | 100 |
| DeepSeek V4.1 Flash 1017 Card - Apex Bench v1.1 | 72.1 | Score (%) | 100 |
| DeepSeek V4.1 Flash 1017 Card - Automation-Bench v1.0.6 | 60.5 | Score (%) | 100 |
| DeepSeek V4.1 Flash 1017 Card - BabyVision (with tools) | 92.5 | Score (%) | 100 |
| DeepSeek V4.1 Flash 1017 Card - Chartography (with tools) | 83 | Score (%) | 100 |
| DeepSeek V4.1 Flash 1017 Card - DeepSWE v1.1 (Mini-SWE-V2 harness) | 77 | Pass@1 (%) | 100 |
| DeepSeek V4.1 Flash 1017 Card - HLE w/o tools (full set) | 42.6 | Accuracy (%) | 100 |
| DeepSeek V4.1 Flash 1017 Card - HLE w/o tools (text-only subset) | 44 | Accuracy (%) | 100 |
| DeepSeek V4.1 Flash 1017 Card - NL2Repo-Bench | 68.2 | Score (%) | 100 |
| DeepSeek V4.1 Flash 1017 Card - ProgramBench | 29.1 | Score (%) | 100 |
| DeepSeek V4.1 Flash 1017 Card - Terminal-Bench 4.0 (Claude Code harness) | 42.3 | Score (%) | 100 |
| DeepSeek V4.1 Flash 1017 Card - ZeroBench-main (with tools, Pass@5) | 51 | Pass@5 (%) | 100 |
Interactive version: theaggregate.ai/model?slug=deepseek-v4-1-flash-1017 · How It Works · Data refreshed daily, snapshot 2026-10-18.