Compare / GPT-5.1 vs Grok 4.6
GPT-5.1 vs Grok 4.6
20 results from 5 sources that measured both models. Each row is in the source's own units; there is no overall winner.
| Benchmark | Metric | GPT-5.1 | Grok 4.6 |
|---|---|---|---|
| Arena (formerly LMArena)reported | OverallArena rating | 1,416 | 1,461 |
| Arena (formerly LMArena)reported | OverallArena rating | 1,459 | 1,449 |
| Arena (formerly LMArena)reported | Business, management and financeArena rating | 1,448 | 1,447 |
| Arena (formerly LMArena)reported | Creative writingArena rating | 1,428 | 1,445 |
| Arena (formerly LMArena)reported | Expert promptsArena rating | 1,482 | 1,490 |
| Arena (formerly LMArena)reported | Instruction followingArena rating | 1,450 | 1,451 |
| Arena (formerly LMArena)reported | OverallArena rating | 1,456 | 1,453 |
| Arena (formerly LMArena)reported | Writing, literature and languageArena rating | 1,437 | 1,446 |
| Artificial Analysisreported | Artificial Analysis Coding Indexindex score | 49.4 | 76.8 |
| Artificial Analysisreported | Artificial Analysis Intelligence Indexindex score | 24.7 | 44.3 |
| Artificial Analysisreported | GPQA Diamond% of questions | 87.3% | 94.9% |
| Artificial Analysisreported | Humanity's Last Exam% of questions | 28.5% | 44.1% |
| Artificial Analysisreported | Long-context reasoning (AA-LCR)% of questions | 80.0% | 81.0% |
| Artificial Analysisreported | Terminal-Bench 2.1% of tasks | 52.4% | 88.4% |
| Artificial Analysisreported | Τ-bench banking% of tasks | 15.9% | 50.7% |
| SimpleQA Verified (Epoch AI)reported | Correct answers% of questions | 48.0% | 49.3% |
| UGI Leaderboardreported | Requested-length error% off the requested word count | 113.0% | 25.0% |
| UGI Leaderboardreported | Style adherencescore from 0 to 1 | 0.38 | 0.36 |
| UGI Leaderboardreported | Writing scorescore out of 100 | 55.8 | 63.3 |
| Vending-Bench 2reported | Money after a yearUS dollars | $1,473.43 | $9,047.03 |
Bold green marks the better value on that metric. Where a source reports ranges that overlap, the difference may not be meaningful; see the benchmark page for ranges.