GPT-5.6 Sol Pro
| Released | 2026-07-09 |
|---|---|
| Price per 1M | Not listed |
| Overall score | 61.9 (early, few results) |
| Provider | OpenAI |
| Weights | Proprietary |
Scores
| Dimension | Score | Rank | Benchmarks |
|---|---|---|---|
| Overall: Balanced | 61.9 | 4 | |
| Overall: Coding | 60.8 | 3 | |
| Overall: Reasoning | 64.1 | 3 | |
| Overall: Agentic work | 60.1 | 3 | |
| Agents | 59.1 | 1 | |
| Reasoning | 64.2 | 2 | |
| Math | 63.8 | 1 |
Configurations
| Configuration | Balanced | Benchmarks |
|---|---|---|
| Unspecified setting | 62.1 ±8.3 | 1 |
| Max | 63.3 ±4.6 | 3 |
| Extra High | 61.6 ±8.3 | 1 |
Evidence
Agents 59.1 ±2.2
| Benchmark | Result | Source | Date |
|---|---|---|---|
| APEX-Agents | 51.4% | Epoch AI | 2026-11-11 |
Reasoning 64.2 ±5.1
| Benchmark | Result | Source | Date |
|---|---|---|---|
| Chess Puzzles | 68.0% | Epoch AI | 2026-10-14 |
| SimpleBench | 71.7% | SimpleBench | 2026-07-09 |
| SimpleBench | 71.7% | Epoch AI | 2026-11-11 |
Math 63.8 ±1.8
| Benchmark | Result | Source | Date |
|---|---|---|---|
| FrontierMath Tier 4 | 80.5% | Epoch AI | 2026-07-09 |
Results not used in the score
| Benchmark | Result | Source | Why not used |
|---|---|---|---|
| SimpleBench | 71.7% | Epoch AI | A stronger configuration result was used |
Names matched to this model
gpt-5-6-sol-pro