Claude Opus 5
Parameters
Undisclosed
Context Length
1M
Modality
Multimodal
Architecture
Undisclosed
License
Proprietary
Release Date
24 Jul 2026
Knowledge Cutoff
May 2026
API Pricing (per 1M)
Input: $5.00 · Output: $25.00
Evaluation Benchmarks
Rank
#3
| Benchmark | Score | Rank |
|---|---|---|
Agent Arena | max 0.12 high 0.14 | 🥈 2 🥇 1 |
Agentic Coding | max 0.65 | 🥈 2 |
Intelligence Index | max 0.54 high 0.52 medium 0.49 low 0.44 | 🥈 2 5 11 33 |
Agentic Index | max 0.56 high 0.53 medium 0.47 low 0.38 | 🥈 2 9 20 38 |
Reasoning | max 0.91 | 🥉 3 |
Web Development | max 1687 high 1662 | 🥉 3 5 |
Coding Index | max 0.78 high 0.77 medium 0.74 low 0.67 | 4 9 21 50 |
General | max 0.80 | 6 |
Mathematics | max 0.96 | 6 |
LiveBench Average | max 0.80 | 6 |
General Text | max 1488 high 1493 | 13 9 |
Coding | max 0.81 | 9 |
Data Analysis | max 0.75 | 23 |
Rankings
Overall Rank
#3 🥉
Coding Rank
#2 🥈
About Claude Opus 5
Claude Opus 5 is Anthropic's flagship model engineered for demanding reasoning, advanced software engineering, and long-horizon agentic workflows. It provides comprehensive visual and document understanding supported by a 1M token context window.
Technical Specifications
Architecture specifications are undisclosed for proprietary models.
About Claude 4
Anthropic's fourth generation Claude models with advanced reasoning, extended context windows up to 200K tokens, and configurable thinking effort levels. Features improved safety alignment, nuanced understanding, and sophisticated task completion. Includes Opus (most capable), Sonnet (balanced), and Haiku (fast) variants, with thinking modes that enable transparent chain-of-thought reasoning for complex problems.