Up to four models
GPT-6 Astra vs GPT-6.1 Sol
On the index (GPT-4 = 100), GPT-6 Astra scores 199 and GPT-6.1 Sol 196. 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 Astra
199182–215
WeightsClosed
ReleasedSep 2026
Price in / out$10 / $50
Context1.05M
Core tests taken7 of 9
GPT-6.1 Sol
196180–212
WeightsClosed
ReleasedSep 2026
Price in / out$2.0 / $10
Context1.05M
Core tests taken6 of 9
GPT-6 Astra and GPT-6.1 Sol are too close to call (3 points apart, within the uncertainty), ahead on 5 of 6 shared core tests, at 5.0× the price.
Test by testthird-party results
GPT-6 AstraGPT-6.1 Sol
Every result as a table
| Test | GPT-6 Astra | GPT-6.1 Sol |
|---|---|---|
| Reasoning & knowledge | ||
| GPQA Diamond (Epoch AI) | 95.8% | 95.4% |
| LiveBench Data Analysis | 83.0% | 82.7% |
| LiveBench IF | 75.6% | 74.2% |
| LiveBench Language | 89.4% | 90.1% |
| LiveBench Reasoning | 92.7% | 92.6% |
| EBR-bench | 76.2% | 54.3% |
| SimpleQA Verified | 75.6% | 73.9% |
| Maths | ||
| FrontierMath (tiers 1-3) | 93.7% | 93.7% |
| AIME-style maths (OTIS mock) | 100.0% | 100.0% |
| LiveBench Mathematics | 96.8% | 96.8% |
| ProofBench v1.1 | 99.0% | 99.0% |
| Chess puzzles | 74.0% | 68.0% |
| FrontierMath (tier 4) | 97.6% | 100.0% |
| Coding | ||
| Terminal-Bench 4.0 | 59.6% | 55.0% |
| LiveBench Agentic Coding | 57.3% | 56.8% |
| LiveBench Coding | 80.4% | 80.7% |
| Code Migration | 67.7% | 65.1% |
| CyberBench v1.1 | 41.1% | 39.3% |
| IOI | 100.0% | 96.9% |
| ProgramBench | 5.5% | 3.0% |
| Vibe Code Bench v1.1 | 89.6% | 88.9% |
| Agents | ||
| APEX-Agents | 64.7% | 60.0% |
| BioMysteryBench | 79.3% | 79.6% |
| SRE Bench | 56.9% | 50.8% |
| Terminal-Bench-Science | 62.9% | 52.9% |
| DTBench | 97.3% | 96.5% |
| Furniture assembly | 80.0% | 80.0% |
| LMCA | 64.4% | 62.8% |
| Mystery game puzzles | 84.0% | 80.0% |
| Professional work | ||
| Finance Agent (v2) | 53.5% | 52.0% |
| EMB | 71.7% | 70.8% |
| HLAB-Vals-1.1 | 7.1% | 6.3% |
| Legal Research Bench | 39.4% | 38.5% |
| MedCode | 48.5% | 48.8% |
| MedScribe | 87.9% | 86.5% |
| SAGE | 46.4% | 46.5% |
| Tax Agent Bench | 20.7% | 21.8% |
| Novel problems | ||
| ARC-AGI-2 | 95.0% | 94.2% |
| ARC-AGI-1 | 98.5% | 98.5% |
| ARC-AGI-3 | 62.7% | 52.7% |
| MysteryMechanism | 53.1% | 46.4% |
Tests at least two of these models have taken. Bold rows are the core tests.
