← All benchmarks
Claude Opus 4.8 (Adaptive Reasoning, Max Effort)
Anthropic · Released 2026-05-28 · Independent benchmark scores from Artificial Analysis
Intelligence Index
55.7
Output speed
66.8
tokens/sec
Blended price
$10.00
per 1M tokens
Time to first token
35.19s
Model metadata
| Model ID | 992b7b84-5069-4c6a-9295-834252553d50 |
|---|---|
| Slug | claude-opus-4-8 |
| Release date | 2026-05-28 |
| Provider | Anthropic |
| Provider slug | anthropic |
| Output speed | 66.8 tok/s |
| Time to first token | 35.19s |
| Time to first answer | 35.19s |
| Blended price ($/1M) | $10.00 |
| Input price ($/1M) | $5.00 |
| Output price ($/1M) | $25.00 |
All evaluation scores (17)
| aime | — |
|---|---|
| aime 25 | — |
| artificial analysis coding index | 74.30 |
| artificial analysis intelligence index | 55.70 |
| artificial analysis math index | — |
| gpqa | 92.0% |
| hle | 45.7% |
| ifbench | 62.2% |
| lcr | 67.7% |
| livecodebench | — |
| math 500 | — |
| mmlu pro | — |
| scicode | 53.5% |
| tau banking | 27.6% |
| tau2 | 94.4% |
| terminalbench hard | 58.3% |
| terminalbench v2 1 | 84.6% |
Source: artificialanalysis.ai/models/claude-opus-4-8
Benchmark data from Artificial Analysis.