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 3 HaikuGPT-4.1 Mini
| Claude 3 Haiku Anthropic | GPT-4.1 Mini OpenAI | |
|---|---|---|
| Pricing | ||
| Input $/1M | $0.25 | $0.40 |
| Output $/1M | $1.25 | $1.60 |
| Blended $/1M (3:1) | $0.50 | $0.70 |
| Cache read $/1M | $0.03 | $0.10 |
| Cache write $/1M | $0.30 | — |
| Quality indices | ||
| Intelligence Index | 3.9 | 14.8 |
| Coding Index | — | 20.2 |
| Math Index | — | 46.3 |
| Agentic Index | — | 1.7 |
| Speed & latency | ||
| Output speed (tok/s) | 0 | 83 |
| Time to first token | 0s | 0.55s |
| Time to first answer token | 0s | 0.55s |
| Benchmarks (Artificial Analysis) | ||
| MMLU-Pro | — | 78.1% |
| GPQA Diamond | 37.4% | 66.4% |
| Humanity's Last Exam | 3.9% | 4.6% |
| LiveCodeBench | 15.4% | 48.3% |
| SciCode | 18.6% | 40.4% |
| MATH-500 | 39.4% | 92.5% |
| AIME | 1.0% | 43.0% |
| AIME 2025 | — | 46.3% |
| IFBench | 36.1% | 38.3% |
| AA-LCR (Long Context Reasoning) | 21.0% | 42.3% |
| Terminal-Bench Hard | 0.8% | 7.6% |
| Terminal-Bench 2.1 | — | 10.1% |
| τ²-Bench (Telecom) | 21.1% | 52.9% |
| τ³-Bench Banking | — | 5.2% |
| Design Arena (Elo) | ||
| Website | — | 1,025 |
| UI components | — | 1,004 |
| Data viz | — | 1,070 |
| 3D | — | 895 |
| Game dev | — | 1,129 |
| Code categories | — | 1,027 |
| Specs | ||
| Context window | 200K | 1.05M |
| Max output tokens | 4K | 33K |
| Input modalities | text, image | image, text, file |
| Tokenizer | Claude | GPT |