Claude Opus 4.8 vs DeepSeek-V4-Pro
Anthropic vs DeepSeek — benchmarks, pricing, and capabilities side by side.
- •DeepSeek-V4-Pro has the higher intelligence index (78.5 vs 73.1)
- •DeepSeek-V4-Pro is cheaper ($0.44 vs $6.30 per 1M input)
- •Claude Opus 4.8 is faster
| Claude Opus 4.8 | DeepSeek-V4-Pro | |
|---|---|---|
| Intelligence index | 73.1 | 78.5 |
| Developer | Anthropic | DeepSeek |
| Type | LLM | LLM |
| Access | — | Open weights |
| Context window | — | 1,048,576 tokens |
| Input price | $6.30 / 1M | $0.44 / 1M |
| Output price | $25.00 / 1M | $0.87 / 1M |
| Speed | 66 tok/s | 30 tok/s |
| Released | May 28, 2026 | April 24, 2026 |
| Parameters | — | 1.6T (49B active) |
| Input modalities | — | Text |
| Output modalities | — | Text |
Shared benchmarks
Claude Opus 4.8
DeepSeek-V4-Pro
AA-LCR
67.7
66.3
GPQA Diamond
92
90.1
Humanity’s Last Exam
45.7
35.9
SciCode
53.5
50
Terminal-Bench
58.3
46.2
τ²-bench
94.4
96.2