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 SolGrok 4.6
| GPT-5.6 Sol OpenAI | Grok 4.6 xAI | |
|---|---|---|
| Pricing | ||
| Input $/1M | $2.00 | $2.00 |
| Output $/1M | $10.00 | $6.00 |
| Blended $/1M (3:1) | $4.00 | $3.00 |
| Cache read $/1M | $0.20 | $0.50 |
| Cache write $/1M | $2.50 | — |
| Quality indices | ||
| Intelligence Index | 47 | 44.3 |
| Coding Index | 77.4 | 76.8 |
| Math Index | — | — |
| Agentic Index | 50.2 | 53 |
| Speed & latency | ||
| Output speed (tok/s) | 65 | 58 |
| Time to first token | 64.46s | 34.50s |
| Time to first answer token | 64.46s | 34.50s |
| Benchmarks (Artificial Analysis) | ||
| GPQA Diamond | 94.1% | 94.9% |
| Humanity's Last Exam | 49.5% | 42.9% |
| SciCode | 57.1% | 56.5% |
| IFBench | 72.7% | — |
| AA-LCR (Long Context Reasoning) | 84.0% | 80.3% |
| Terminal-Bench Hard | 65.9% | — |
| Terminal-Bench 2.1 | 88.0% | 88.4% |
| τ²-Bench (Telecom) | 85.1% | — |
| τ³-Bench Banking | 44.3% | 50.7% |
| Design Arena (Elo) | ||
| Website | — | 1,304 |
| Web apps | — | 1,264 |
| Full-stack | — | 1,274 |
| Mobile apps | — | 1,267 |
| Android native | — | 1,301 |
| UI components | — | 1,305 |
| Data viz | — | 1,303 |
| SVG | — | 1,263 |
| 3D | — | 1,305 |
| Game dev | — | 1,321 |
| Agentic game dev | — | 1,210 |
| HTML slides | — | 1,253 |
| Python→PPTX slides | — | 1,242 |
| Code categories | — | 1,309 |
| ASCII art | — | 1,297 |
| Specs | ||
| Context window | 1.05M | 500K |
| Max output tokens | 128K | 450K |
| Input modalities | file, image, text | text, image, file |
| Tokenizer | GPT | Grok |