Up to four models
GPT-6 Sol vs Hy4 preview
On the index (GPT-4 = 100), GPT-6 Sol scores 178 and Hy4 preview 173. 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 Sol
178162–194
WeightsClosed
ReleasedSep 2026
Price in / out$2.0 / $10
Context1.05M
Core tests taken6 of 9
173154–192
WeightsOpen
ReleasedAug 2026
Price in / out$0.83 / $2.5
Context1.05M
Core tests taken2 of 9
GPT-6 Sol and Hy4 preview are too close to call (5 points apart, within the uncertainty), ahead on 1 of 2 shared core tests, at 3.2× the price.
Test by testthird-party results
GPT-6 Sol
Hy4 preview
Every result as a table
| Test | GPT-6 Sol | Hy4 preview |
|---|---|---|
| Maths | ||
| ProofBench v1.1 | 83.0% | 75.0% |
| Coding | ||
| Terminal-Bench 4.0 | 44.4% | 39.4% |
| Code Migration | 57.2% | 47.4% |
| CyberBench v1.1 | 78.0% | 65.4% |
| IOI | 82.6% | 59.3% |
| ProgramBench | 2.0% | 0.0% |
| Vibe Code Bench v1.1 | 87.8% | 81.0% |
| Agents | ||
| BioMysteryBench | 74.8% | 69.3% |
| SRE Bench | 30.5% | 2.3% |
| Professional work | ||
| Finance Agent (v2) | 49.0% | 55.1% |
| EMB | 71.5% | 58.1% |
| Legal Research Bench | 28.8% | 45.2% |
| MedCode | 47.1% | 43.3% |
| MedScribe | 82.0% | 83.6% |
| Public Benefits Bench v1.1 | 56.6% | 68.6% |
| Tax Agent Bench | 15.0% | 34.4% |
Tests at least two of these models have taken. Bold rows are the core tests.