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.
Claude Opus 4.5GPT-5.6 Terra
| Claude Opus 4.5 Anthropic | GPT-5.6 Terra OpenAI | |
|---|---|---|
| Pricing | ||
| Input $/1M | $5.00 | $2.50 |
| Output $/1M | $25.00 | $15.00 |
| Blended $/1M (3:1) | $10.00 | $5.63 |
| Cache read $/1M | $0.50 | $0.25 |
| Cache write $/1M | $6.25 | $3.13 |
| Quality indices | ||
| Intelligence Index | 34.7 | 55 |
| Coding Index | — | 76.7 |
| Math Index | 62.7 | — |
| Agentic Index | — | 47.4 |
| Speed & latency | ||
| Output speed (tok/s) | 68 | 149 |
| Time to first token | 1.30s | 104.73s |
| Time to first answer token | 1.30s | 104.73s |
| Benchmarks (Artificial Analysis) | ||
| MMLU-Pro | 88.9% | — |
| GPQA Diamond | 81.0% | 92.5% |
| Humanity's Last Exam | 12.9% | 41.8% |
| LiveCodeBench | 73.8% | — |
| SciCode | 47.0% | 53.9% |
| AIME 2025 | 62.7% | — |
| IFBench | 43.0% | 71.2% |
| AA-LCR (Long Context Reasoning) | 65.3% | 74.0% |
| Terminal-Bench Hard | 40.9% | 57.6% |
| Terminal-Bench 2.1 | — | 88.0% |
| τ²-Bench (Telecom) | 86.3% | 86.3% |
| τ³-Bench Banking | — | 31.8% |
| Design Arena (Elo) | ||
| Website | 1,274 | — |
| Web apps | 1,216 | — |
| Full-stack | 1,215 | — |
| Mobile apps | 1,254 | — |
| Android native | 1,169 | — |
| UI components | 1,280 | — |
| Data viz | 1,278 | — |
| SVG | 1,227 | — |
| 3D | 1,272 | — |
| Game dev | 1,288 | — |
| Code categories | 1,276 | — |
| ASCII art | 1,231 | — |
| Specs | ||
| Context window | 200K | 1.05M |
| Max output tokens | 64K | 128K |
| Input modalities | file, image, text | file, image, text |
| Tokenizer | Claude | GPT |