Side-by-side comparison of AI model specs, benchmarks, and pricing.
| Spec | Claude Haiku 4.5 Anthropic | Grok 4.1 xAI |
|---|---|---|
| Tier | mid | frontier |
| Release Date | 2025-10-01 | 2025-11-17 |
| Context Window | 200K | 2000K |
| Max Output | 64K | 16K |
| Input Price (/1M) | $1 | $0.2 |
| Output Price (/1M) | $5 | $0.5 |
| Arena Elo | 1,404 | 1,465 |
| MMLU | 82% | 90.2% |
| GPQA | 55% | 88% |
| MATH | — | — |
| HumanEval | 90% | 92% |
| SWE-bench | — | 65% |
| Capabilities | vision, tool-use, code, extended-thinking | vision, tool-use, code, reasoning |