Claude Opus 4.7
ClaudeBalanced
ThinkingTool UseVisionStructured Output
About this model
Opus 4.7 is the next generation of Anthropic's Opus family, built for long-running, asynchronous agents. Building on the coding and agentic strengths of Opus 4.6, it delivers stronger performance on...
Performance Tier
Balanced
Claude Opus 4.7 is a balanced model from Claude : strong performance at a reasonable price.
Strong cost-performance ratio. Reliable for most professional use cases without premium pricing.
Pricing
This model is included in Elosia plansPremium
Highest cost level. A long conversation can quickly consume your monthly cap.
| Type | per 1M tokens |
|---|---|
| Input (prompt) | $5.00 |
| Output (completion) | $25.00 |
| Cache read | $0.500 |
| Cache write | $6.25 |
Capabilities
Context Length1.0M
Max Output Tokens128K
TokenizerClaude
Inputtext, image, file
Outputtext
Release DateApril 16, 2026
Benchmarks
General Intelligence
MMLU
91.5%
GPQA Diamond
94.2%
Mathematics
MATH-500
Not reported
Programming
HumanEval
Not reported
SWE-bench Verified
87.6%
Reasoning
ARC-AGI-2
75.8%
Humanity's Last Exam
46.9%
Agentic
SWE-bench Pro
64.3%
Terminal-Bench 2.0
69.4%
Where does Claude Opus 4.7 stand?
Compare its performance index against every other model.
Recommended Use Cases
CodingAnalysisResearchCreative Writing
Strengths
- Best generally-available model on real-world software engineering (SWE-bench Verified 87.6%, SWE-bench Pro 64.3%)
- State-of-the-art graduate-level scientific reasoning (GPQA Diamond 94.2%)
- Adaptive thinking — automatically adjusts compute to task complexity
- Improved agentic robustness against prompt injection and malicious tool use
- Enhanced vision: accepts images up to ~3.75 MP (3× prior Claude models)
Limitations
- Premium pricing ($5 / $25 per million input / output tokens)
- MATH-500, HumanEval and IFEval not reported — Anthropic considers them saturated
- Slower than Sonnet/Haiku for latency-sensitive queries
Frequently asked questions
Resources
This model may use your data for training