Up to four models
GPT-6 vs DeepSeek V4.1 Flash 1017
On the index (GPT-4 = 100), GPT-6 scores 199 and DeepSeek V4.1 Flash 1017 176. Side by side on the same third-party tests, with price and context. The link in your address bar shares this exact comparison.
DeepSeek V4.1 Flash 1017176160–192
WeightsOpen
ReleasedOct 2026
Price in / out$0.30 / $1.2
Context1.05M
Core tests taken5 of 9
GPT-6 leads DeepSeek V4.1 Flash 1017 by 23 points, ahead on 4 of 5 shared core tests.
Test by testthird-party results
GPT-6DeepSeek V4.1 Flash 1017
DeepSeek V4.1 Flash 1017Every result as a table
| Test | GPT-6 | DeepSeek V4.1 Flash 1017 |
|---|---|---|
| Reasoning & knowledge | ||
| GPQA Diamond (Epoch AI) | 95.8% | 93.2% |
| SimpleQA Verified | 75.6% | 42.5% |
| Maths | ||
| FrontierMath (tiers 1-3) | 93.7% | 71.2% |
| AIME-style maths (OTIS mock) | 100.0% | 99.4% |
| Chess puzzles | 74.0% | 46.0% |
| FrontierMath (tier 4) | 97.6% | 46.3% |
| Coding | ||
| Terminal-Bench 4.0 | 59.6% | 31.8% |
| Code Migration | 67.7% | 44.2% |
| Vibe Code Bench v1.1 | 89.6% | 87.7% |
| Agents | ||
| Mystery game puzzles | 84.0% | 49.0% |
| Professional work | ||
| Finance Agent (v2) | 53.5% | 57.0% |
| EMB | 71.7% | 61.8% |
| HLAB-Vals-1.1 | 7.1% | 14.2% |
| Legal Research Bench | 39.4% | 45.2% |
| Tax Agent Bench | 20.7% | 44.4% |
| Novel problems | ||
| ARC-AGI-2 | 95.0% | 78.1% |
| ARC-AGI-1 | 98.5% | 94.5% |
Tests at least two of these models have taken. Bold rows are the core tests.
