Up to four models
GPT-6.1 Sol vs GPT-6 Sol
On the index (GPT-4 = 100), GPT-6.1 Sol scores 196 and GPT-6 Sol 185. Side by side on the same third-party tests, with price and context. The link in your address bar shares this exact comparison.
GPT-6.1 Sol
196180–212
WeightsClosed
ReleasedSep 2026
Price in / out$2.0 / $10
Context1.05M
Core tests taken6 of 9
GPT-6 Sol
185169–201
WeightsClosed
ReleasedSep 2026
Price in / out$2.0 / $10
Context1.05M
Core tests taken6 of 9
GPT-6.1 Sol leads GPT-6 Sol by 11 points, ahead on 6 of 6 shared core tests, at a similar price.
Test by testthird-party results
GPT-6.1 SolGPT-6 Sol
Every result as a table
| Test | GPT-6.1 Sol | GPT-6 Sol |
|---|---|---|
| Reasoning & knowledge | ||
| GPQA Diamond (Epoch AI) | 95.4% | 94.3% |
| LiveBench Data Analysis | 82.7% | 81.2% |
| LiveBench IF | 74.2% | 68.6% |
| LiveBench Language | 90.1% | 85.3% |
| LiveBench Reasoning | 92.6% | 88.6% |
| EBR-bench | 54.3% | 53.3% |
| SimpleQA Verified | 73.9% | 60.7% |
| Maths | ||
| FrontierMath (tiers 1-3) | 93.7% | 89.8% |
| AIME-style maths (OTIS mock) | 100.0% | 100.0% |
| LiveBench Mathematics | 96.8% | 96.4% |
| ProofBench v1.1 | 99.0% | 83.0% |
| Chess puzzles | 68.0% | 56.0% |
| FrontierMath (tier 4) | 100.0% | 90.0% |
| Coding | ||
| Terminal-Bench 4.0 | 55.0% | 44.4% |
| LiveBench Agentic Coding | 56.8% | 52.9% |
| LiveBench Coding | 80.7% | 81.8% |
| Code Migration | 65.1% | 57.2% |
| CyberBench v1.1 | 39.3% | 78.0% |
| IOI | 96.9% | 82.6% |
| ProgramBench | 3.0% | 2.0% |
| Vibe Code Bench v1.1 | 88.9% | 87.8% |
| Agents | ||
| APEX-Agents | 60.0% | 54.3% |
| BioMysteryBench | 79.6% | 74.8% |
| SRE Bench | 50.8% | 30.5% |
| Terminal-Bench-Science | 52.9% | 30.0% |
| DTBench | 96.5% | 97.3% |
| Furniture assembly | 80.0% | 58.3% |
| LMCA | 62.8% | 59.1% |
| Mystery game puzzles | 80.0% | 56.0% |
| Professional work | ||
| Finance Agent (v2) | 52.0% | 49.0% |
| EMB | 70.8% | 71.5% |
| Legal Research Bench | 38.5% | 28.8% |
| MedCode | 48.8% | 47.1% |
| MedScribe | 86.5% | 82.0% |
| Public Benefits Bench v1.1 | 59.3% | 56.6% |
| SAGE | 46.5% | 44.8% |
| Tax Agent Bench | 21.8% | 15.0% |
| Novel problems | ||
| ARC-AGI-2 | 94.2% | 89.6% |
| ARC-AGI-1 | 98.5% | 95.5% |
| ARC-AGI-3 | 52.7% | 4.6% |
| MysteryMechanism | 46.4% | 30.2% |
Tests at least two of these models have taken. Bold rows are the core tests.
