LLM

Claude Opus 4.7 (Adaptive Reasoning, Max Effort)

Name: Claude Opus 4.7 (Adaptive Reasoning, Max Effort) Review & Benchmarks
Item: Claude Opus 4.7 (Adaptive Reasoning, Max Effort)
Rating: 57.3
Author: Sentient Weekly

by Anthropic

Overall Score

57.3

Anthropic's flagship reasoning model

Released Apr 2026

Benchmark Scores

Reasoning

—

Coding

52.5

Math

—

Creative writing

—

Instruction following

—

Multimodal

—

Standard Benchmarks

aa_lcraa_lcr

0.7

Measured 2026-05-13 · source

aa_terminalbench_hardaa_terminalbench_hard

0.5

Measured 2026-05-13 · source

aa_gpqaaa_gpqa

0.9

Measured 2026-05-13 · source

aa_hleaa_hle

0.4

Measured 2026-05-13 · source

aa_scicodeaa_scicode

0.6

Measured 2026-05-13 · source

aa_ifbenchaa_ifbench

0.6

Measured 2026-05-13 · source

aa_tau2aa_tau2

0.9

Measured 2026-05-13 · source

Price

$15 / 1M input · $75 / 1M output

Paid

Context

1M tokens

Speed

71.62 t/s

20608ms TTFT

Compare this model →

Overview

Anthropic's flagship reasoning model. Excellent at complex multi-step analysis, long-context retrieval, agentic workflows, and producing high-quality writing. The 1M-token context window enables full-codebase or full-document workflows.

Strengths

State-of-the-art reasoning · 1M-token context · best-in-class agentic tool use · strong creative writing · careful, calibrated answers

Known limitations

Higher cost per token than smaller models · slightly slower than Sonnet · still has some refusal/over-cautiousness on legitimate edge cases

Discussion

0 comments

Be the first to comment.