Atlas

Models

← All models

Sonnet 4.6

Anthropic · 2026-02-17

Sonnet 4.6 is Anthropic's balanced mid-tier Claude model, released as a hybrid-reasoning model supporting both extended and adaptive thinking. It improved broadly over Sonnet 4.5 across coding, computer use, long-context reasoning, agent planning, and knowledge work, with a 1M-token context window.

Benchmark scores

BenchmarkScore
A-Fantasia52.3
A-Fantasia (Valid Attempts with Retries)47.7
A-Fantasia Backwards Spelling39.0
A-Fantasia Chess26.0
A-Fantasia Chess (Valid Attempts with Retries)30.0
A-Fantasia Cube (Valid Attempts with Retries)73.0
A-Fantasia Cube Rotation92.0
A-Fantasia Spelling (Valid Attempts with Retries)40.0
AA-AnalystAgent (pass@1)46.8
AA-AnalystAgent (pass@5)71.3
AA-AnalystAgent (pass^5)20.0
AA-Briefcase Elo1065
AA-Briefcase v1.1 Elo1062
AA-LCR70.7
AA-LCR v1.180.5
AA-Omniscience Index12.2
AdvancedIF70.8
AdvancedIF (rubric mean score)92.4
Agent Red Teaming k=100 (CAIS)54.8
Agent4Sci mean relative improvement9.6
Agents' Last Exam ALE-CLI Pass Rate13.3
Agents' Last Exam ALE-CLI Score32.0
AHK-Eval97.2
AHK-Eval Algorithms66.7
AHK-Eval Data Structures50.0
AHK-Eval Date and Time100.0
AHK-Eval Easy75.0
AHK-Eval Hard58.3
AHK-Eval Hidden Cases99.4
AHK-Eval Mid75.0
Loading Atlas data…