Compare models
Pick up to 4 models for a side-by-side spec sheet. Hover any dotted metric name for what the test measures. Best value in each row is highlighted.
GPT-5.6 SolHy3 preview
| GPT-5.6 Sol OpenAI | Hy3 preview Tencent | |
|---|---|---|
| Pricing | ||
| Input $/1M | $5.00 | $0.063 |
| Output $/1M | $30.00 | $0.21 |
| Blended $/1M (3:1) | $11.25 | $0.10 |
| Cache read $/1M | $0.50 | $0.021 |
| Cache write $/1M | $6.25 | — |
| Quality indices | ||
| Intelligence Index | 58.9 | 33.6 |
| Coding Index | 77.4 | 58.8 |
| Math Index | — | — |
| Agentic Index | 54 | 30.7 |
| Speed & latency | ||
| Output speed (tok/s) | 70 | 150 |
| Time to first token | 112.90s | 1.68s |
| Time to first answer token | 112.90s | 15.05s |
| Benchmarks (Artificial Analysis) | ||
| GPQA Diamond | 94.1% | 86.7% |
| Humanity's Last Exam | 47.2% | 25.5% |
| SciCode | 56.1% | 41.2% |
| IFBench | 72.7% | 63.1% |
| AA-LCR (Long Context Reasoning) | 73.7% | 54.7% |
| Terminal-Bench Hard | 65.9% | 34.1% |
| Terminal-Bench 2.1 | 88.0% | — |
| τ²-Bench (Telecom) | 85.1% | 92.7% |
| τ³-Bench Banking | 33.0% | — |
| Specs | ||
| Context window | 1.05M | 262K |
| Max output tokens | 128K | — |
| Input modalities | file, image, text | text |
| Tokenizer | GPT | Other |