Claude Opus (Anthropic)
Claude Opus by Anthropic scores highest on reasoning (98/100) in Nivana’s weekly evals. Price: $4 in / $20 out per 1M tokens. Context window: 1,000,000 tokens.
Nivana’s router picks Claude Opus first for: Hard reasoning.
Quality by task
| Task | Score (0–100) |
|---|---|
| reasoning and analysis | 98 |
| coding | 97 |
| writing | 95 |
| maths | 93 |
| translation | 91 |
| long documents | 91 |
| structured data and JSON | 88 |
| quick answers and everyday questions | 83 |
Compare: Claude Opus vs GPT-6.1 Sol · Claude Opus vs Gemini 4 Argon · Claude Sonnet vs Claude Opus · DeepSeek V4 Pro vs Claude Opus