Claude Opus 4.7

Mid-tier

Claude Opus 4.7 was Anthropic's flagship from April 17, 2026 until Opus 4.8 succeeded it on May 28, 2026. It introduced a 1 million token context window at the same API pricing as 4.6, alongside gains on reasoning, math, and SWE-bench. It remains available on the API at $15 input and $75 output per million tokens, and is a solid choice for long-context code work, multi-document analysis, and agent workflows that span hours of tool calls.

Input Price

$15.00

per 1M tokens

Output Price

$75.00

per 1M tokens

Context Window

tokens

Released

2026-04

API access

Capabilities

textvisiontool-usecode

Key Strengths

✓1M token context window
✓Leads on HumanEval and SWE-bench
✓Same price as 4.6 with 5x context
✓Strong agentic tool use

Best For

▸Whole-repository refactors
▸Multi-document research synthesis
▸Long-running agent workflows
▸Extended codebase debugging

Benchmark Scores

Benchmark	Score	Description
MMLU-Pro	93.8	General knowledge and reasoning across 57 subjects
HumanEval	96.2	Python code generation and problem solving
GPQA Diamond	76.5	Graduate-level science questions verified by domain experts
MATH	93.1	Competition-level mathematics problems
SWE-bench	65.4	Real-world software engineering tasks from GitHub issues