Claude Opus 4.6

Frontier

Anthropic•Released on 2026-02-05

Anthropic's flagship model with 1M token context (beta), adaptive thinking, and the highest agentic coding scores. Introduced Agent Teams for parallel autonomous coding. Nearly doubled ARC-AGI-2 score over Opus 4.5 (68.8% vs 37.6%).

91
Overall Score

Core Specs

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

Scenario Scores

Coding
#1 / 1496
Writing
#3 / 1491
Translation
#4 / 1489
Data Analysis
#2 / 1490
Conversation
#3 / 1488
Image Gen
—

User Feedback Highlights

Based on community feedback. Hover to see original reviews.

+ Highest SWE-bench score (80.8%)+ Agent Teams for parallel coding− Response prefilling removed (breaking change)+ 128K max output (doubled from 4.5)− 2x price of GPT-5.4+ Adaptive thinking with effort levels+ Best instruction following in complex contexts− 1M context in beta only− Extended thinking deprecated
Sentiment:👍 78%😐 14%👎 8%

Pros & Cons

Pros

  • +Highest SWE-bench score (80.8%)
  • +128K max output (doubled from 4.5)
  • +Adaptive thinking with effort levels
  • +Agent Teams for parallel coding
  • +Best instruction following in complex contexts

Cons

  • −2x price of GPT-5.4
  • −Response prefilling removed (breaking change)
  • −1M context in beta only
  • −Extended thinking deprecated

Reliability

SLA99.9%
Incidents (30d)4
Last Incident2026-03-05
View Status Page →

Pricing

Input (per 1M tokens)$5.00
Output (per 1M tokens)$25.00
Subscription$20/month
Updated on 2026-03-06

Tools Supporting This Model

Claude CodeDefault
Anthropic
Cline
Open Source
Cursor
Anysphere

Compare with Others

Claude Opus 4.6 vs GPT-5.4 →Claude Opus 4.6 vs GPT-5.4 Thinking →Claude Opus 4.6 vs Claude Opus 4.5 →

Benchmarks

sweBenchVerified80.8%
terminalBench265.4%
browseComp84%
gdpvalElo1606%
arcAgi268.8%
gpqaDiamond91.3%
bigLawBench90.2%