MODELS
Claude Opus 4.7
Anthropic's top-tier reasoning model for long-context, agentic, and code-heavy work.
Specs
- Context window
- 1,000,000
- Max output
- 32,000
- Modalities
- text, image
- Tool use
- ✓
- Vision
- ✓
- Streaming
- ✓
- License
- proprietary
- Released
- 2026-04-01
Pricing
- Input / 1M
- $15.00
- Output / 1M
- $75.00
- Cached input / 1M
- $1.50
Cost estimate
Claude Opus 4.7 is Anthropic's flagship in the Claude 4 line, with a 1M-token context window and 32K max output. It handles text and image input, supports tool use and streaming, and is positioned for hard reasoning, multi-step agents, and large codebase work. Aimed at teams who need depth over speed and are willing to pay premium rates for fewer retries.
Editor's verdict
Pick Opus 4.7 when the task is genuinely hard — multi-file refactors, long agent traces, dense document review — and a wrong answer costs more than the API bill. At $15/$75 per million tokens it's roughly 5× the price of Sonnet-tier models and ~15× GPT-5-mini, so it's a poor default for chat or boilerplate. Closed weights and Anthropic-only hosting also rule it out for on-prem needs.
Reviews
No reviews yet. Be the first.
Last updated: 2026-04-29