Claude Opus 4.6
Released 2026-02 · hybrid · 1.0M tokens · 8 benchmarks
Editorial notes
Lanzado 5 febrero 2026. 1M context GA el 13 marzo. #1 LMArena overall (1504 ELO). Lidera Terminal-Bench 2.0 y Humanity's Last Exam.
Spec sheet
- Company
- Anthropic
- Country
- US
- Type
- hybrid
- Release
- 2026-02
- Context
- 1.0M tokens
- Pricing (anthropic)
- $15/$75/M
- Slug
- claude-opus-4-6
Benchmarks (8)
Reasoning 5
- 92.1MMLUMassive Multitask Language Understanding — 57 academic subjects, ~16K questions.
- 91.3GPQA-DiamondGraduate-level Physics, Chemistry, Biology — PhD-level questions.
- 90.5MMLU-ProMMLU upgraded with harder questions and 10 answer options.
- 68.8ARC-AGI-2Updated ARC challenge — novel and tough abstract reasoning.
- 40.0Humanitys-Last-ExamThe hardest known benchmark — novel academic problems.
Coding 2
Cite this model
BibTeX · APA
BibTeX
@misc{frontier-claude-opus-4-6,
title = {Claude Opus 4.6},
author = {{Anthropic}},
year = {2026},
note = {Frontier Benchmarks AI atlas. Accessed 2026-05-08},
url = {https://frontierbenchmarks.com/models/claude-opus-4-6}
} APA
Anthropic (2026). Claude Opus 4.6 [Large language model]. Frontier Benchmarks AI. Retrieved 2026-05-08, from https://frontierbenchmarks.com/models/claude-opus-4-6
Citation reflects the atlas page, not the original model paper. For the paper, see the "Resources" section above.