Claude Opus 5 (batch)
Anthropic · Multimodal · Released Jul 2026
Claude Opus 5 is Anthropic's flagship reasoning model, designed for complex multi-step problems, software development, and extended agentic work across code review, bug detection, and visual analysis.
- Strengths
- Strong at end-to-end software tasks, deep reasoning, and maintaining context over long interactions with a 1M token window.
- Best for
- Demanding reasoning work, software development projects, code review and debugging, and agentic systems that require sustained reasoning over extended context.
- Limitations
- Newer than Opus 4.8 and positioned as the current Opus-tier flagship, but Anthropic has also released Claude Fable 5, a higher-tier Mythos-class model for the most demanding reasoning and agentic work.
Input / 1M
$2.50
Output / 1M
$12.50
Cached input / 1M
$0.25
Context window
1M
Also billed
- Cache write
- $3.12 / 1M
Price history
Snapshots
| Effective | Input | Output | Cached in | Note | Source |
|---|---|---|---|---|---|
| 28 Jul 2026 | $2.50 | $12.50 | $0.25 | Imported from OpenRouter | openrouter.ai |