Anthropic

Claude Sonnet 4.6

Multimodal
New
#1τ2-bench Telecom
#1τ2-bench Retail
#1MMMLU
+11 more

by Anthropic

+
+
+
+
About

Claude Sonnet 4.6, released by Anthropic in February 2026, is a comprehensive upgrade across coding, computer use, long-reasoning, and agent planning. Released alongside Claude Opus 4.6, it continued Anthropic's pattern of iterating quickly on the Sonnet line to maintain performance for the agentic coding workflows that Claude Code and similar tools rely on day-to-day.

+
+
+
+
Pricing Range
Input (per 1M)$3.00 -$3.00
Output (per 1M)$15.00 -$15.00
Providers3
+
+
+
+
Timeline
AnnouncedFeb 17, 2026
ReleasedFeb 17, 2026
Knowledge CutoffAug 1, 2025
+
+
+
+
Specifications
Capabilities
Multimodal
+
+
+
+
License & Family
License
Proprietary
Performance Overview
Performance metrics and category breakdown

Overall Performance

16 benchmarks
Average Score
72.0%
Best Score
97.9%
High Performers (80%+)
5

Performance Metrics

Max Context Window
264.0K

Top Categories

Science
89.9%
Agents
73.6%
Coding
69.3%
Finance
63.3%
Tool Use
61.3%
+
+
+
+
All Benchmark Results for Claude Sonnet 4.6
Complete list of benchmark scores with detailed information
τ2-bench Telecom97.90
97.9%
Unverified
τ2-bench Retail91.70
91.7%
Unverified
GPQA Diamond
Science
89.90
89.9%
Unverified
MMMLU89.30
89.3%
Unverified
GDPVal-AA Elo1633.00
81.7%
Unverified
SWE Bench Verified
Coding
79.60
79.6%
Unverified
MMMU-Pro (with tools)75.60
75.6%
Unverified
BrowseComp
Agents
74.70
74.7%
Unverified
MMMU-Pro74.50
74.5%
Unverified
OSWorld
Agents
72.50
72.5%
Unverified
Showing 1 to 10 of 16 benchmarks