Claude Sonnet 4.6

Frontier

Anthropic•Released on 2026-02-17

Anthropic's most capable Sonnet yet. 1M context window (beta), 30-50% faster than Sonnet 4.5, approaching Opus-level intelligence at 1/3 the cost. Default model on claude.ai. Excels at coding, computer use, agent planning, and long-context reasoning.

88
Overall Score

Core Specs

1000K
Context Window
128K
Max Output
✓
Reasoning
✗
Open Source
Multimodal Support
textimage

Scenario Scores

Coding
#4 / 2493
Writing
#7 / 2386
Translation
#9 / 2386
Data Analysis
#7 / 2487
Conversation
#5 / 2387
Image Gen
—

User Feedback Highlights

Based on community feedback. Hover to see original reviews.

− Some users report inconsistent reasoning+ Excellent for agentic workflows+ 1M context window (beta)+ Human-level computer use capability+ Near-Opus intelligence at Sonnet price− 1M context still in beta− Still not Opus for most complex tasks+ 30-50% faster than Sonnet 4.5
Sentiment:👍 78%😐 14%👎 8%

Pros & Cons

Pros

  • +1M context window (beta)
  • +Near-Opus intelligence at Sonnet price
  • +30-50% faster than Sonnet 4.5
  • +Human-level computer use capability
  • +Excellent for agentic workflows

Cons

  • −1M context still in beta
  • −Some users report inconsistent reasoning
  • −Still not Opus for most complex tasks

Reliability

SLA99.9%
Incidents (30d)1
与 Opus 共享基础设施
View Status Page →

Pricing

Input (per 1M tokens)$3.00
Output (per 1M tokens)$15.00
Subscription$20/month
Free trial available
Updated on 2026-03-07

Compare with Others

Claude Sonnet 4.6 vs GPT-5.4 →Claude Sonnet 4.6 vs GPT-5.4 Thinking →Claude Sonnet 4.6 vs Claude Opus 4.6 →

Benchmarks

sweBench80%
osWorld72%
mmlu91.2%
humanEval95.5%