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.
Llama 4 ScoutGPT-5.6 Sol
| Llama 4 Scout Meta | GPT-5.6 Sol OpenAI | |
|---|---|---|
| Pricing | ||
| Input $/1M | $0.10 | $5.00 |
| Output $/1M | $0.30 | $30.00 |
| Blended $/1M (3:1) | $0.15 | $11.25 |
| Cache read $/1M | — | $0.50 |
| Cache write $/1M | — | $6.25 |
| Quality indices | ||
| Intelligence Index | 10 | 58.9 |
| Coding Index | 8.2 | 77.4 |
| Math Index | 14 | — |
| Agentic Index | 1.1 | 54 |
| Speed & latency | ||
| Output speed (tok/s) | 76 | 67 |
| Time to first token | 0.56s | 120.38s |
| Time to first answer token | 0.56s | 120.38s |
| Benchmarks (Artificial Analysis) | ||
| MMLU-Pro | 75.2% | — |
| GPQA Diamond | 58.7% | 94.1% |
| Humanity's Last Exam | 4.3% | 47.2% |
| LiveCodeBench | 29.9% | — |
| SciCode | 17.0% | 56.1% |
| MATH-500 | 84.4% | — |
| AIME | 28.3% | — |
| AIME 2025 | 14.0% | — |
| IFBench | 39.5% | 72.7% |
| AA-LCR (Long Context Reasoning) | 25.8% | 73.7% |
| Terminal-Bench Hard | 1.5% | 65.9% |
| Terminal-Bench 2.1 | 3.7% | 88.0% |
| τ²-Bench (Telecom) | 15.5% | 85.1% |
| τ³-Bench Banking | 3.3% | 33.0% |
| Design Arena (Elo) | ||
| Website | 778 | — |
| UI components | 811 | — |
| Data viz | 933 | — |
| Game dev | 831 | — |
| Code categories | 823 | — |
| Specs | ||
| Context window | 1.31M | 1.05M |
| Max output tokens | 16K | 128K |
| Input modalities | text, image | file, image, text |
| Tokenizer | Llama4 | GPT |