Claude Opus 4.5 vs Claude Opus 4.8
Anthropic vs Anthropic — benchmarks, pricing, and capabilities side by side.
- •Claude Opus 4.5 has the higher intelligence index (76.8 vs 73.1)
- •Claude Opus 4.5 is cheaper ($5.00 vs $6.30 per 1M input)
- •Claude Opus 4.8 is faster
| Claude Opus 4.5 | Claude Opus 4.8 | |
|---|---|---|
| Intelligence index | 76.8 | 73.1 |
| Developer | Anthropic | Anthropic |
| Type | LLM | LLM |
| Access | API only | — |
| Context window | 200,000 tokens | — |
| Input price | $5.00 / 1M | $6.30 / 1M |
| Output price | $25.00 / 1M | $25.00 / 1M |
| Speed | 58 tok/s | 66 tok/s |
| Released | November 24, 2025 | May 28, 2026 |
| Parameters | — | — |
| Input modalities | Text, Image | — |
| Output modalities | Text | — |
Shared benchmarks
Claude Opus 4.5
Claude Opus 4.8
AA-LCR
74
67.7
GPQA Diamond
87
92
Humanity’s Last Exam
28.4
45.7
SciCode
49.5
53.5
Terminal-Bench
47
58.3
τ²-bench
89.5
94.4