Core Specifications
- Developer: Anthropic
- Model Family: Claude 4.6 Series
- Primary Models: Claude Opus 4.6 & Claude Sonnet 4.6
- Release Dates: Opus 4.6 — February 5, 2026; Sonnet 4.6 — February 17, 2026
- Model Purpose: Frontier reasoning (Opus) and versatile reasoning + production workflows (Sonnet)
Technical Specifications
- API Model IDs:
claude-opus-4-6, claude-sonnet-4-6 ([platform docs](https://platform.claude.com/docs/en/about-claude/models/whats-new-claude-4-6))
- Context Window: 200K tokens (standard), up to 1M tokens (beta on API)
- Maximum Output Tokens: Opus 4.6 — up to 128K; Sonnet 4.6 — ~64K
- Thinking Modes: Adaptive thinking, extended thinking, and context compaction features
Performance & Capabilities
- Opus 4.6: Flagship intelligence for deep reasoning, agentic planning, and enterprise coding (state-of-the-art benchmark results across coding and agents) ([Anthropic news](https://www.anthropic.com/news/claude-opus-4-6))
- Sonnet 4.6: Balanced, high-performance model with strong coding, computer use, professional reasoning, and extended workflows at lower cost than Opus ([Anthropic news](https://www.anthropic.com/news/claude-sonnet-4-6))
- Long-context Reasoning: Both models support very large context inputs, enabling processing of entire projects, documents, and codebases.
- Benchmarks: Opus 4.6 leads on advanced agentic coding benchmarks; Sonnet 4.6 delivers near-Opus performance on many real-world tasks at mid-tier cost.
Use Cases
- Opus 4.6: Complex codebase refactoring, financial/legal analysis, multicore agent workflows, enterprise automation
- Sonnet 4.6: Coding, spreadsheet automation, browser tasks, knowledge workflows, content generation, tool integration
- Shared Abilities: Support for autonomous workflows, tool use, long reasoning sessions, and professional content creation
Pricing & Availability
- Sonnet 4.6 Pricing: ~$3 per million input tokens and $15 per million output tokens (same as predecessor) ([Anthropic docs](https://www.anthropic.com/claude/sonnet))
- Opus 4.6 Pricing: ~$5 per million input tokens and $25 per million output tokens
- Access: Available via Claude.ai, Claude Cowork, Claude Code, and the Claude API; supported on major cloud platforms
- Free & Pro Tiers: Sonnet 4.6 is default on Free and Pro plans
Multimodal & Tool Support
- Support for coding tasks, tool calling, and computer/browser automation
- Extended reasoning with search, web-fetch, memory, programmatic tool calls
- Context compaction to maintain coherence over long tasks
Limitations & Notes
- 1M token context window currently in beta on API — default standard limits remain lower
- Pricing may vary by cloud provider and region
- Sonnet 4.6 trades peak frontier power for cost efficiency relative to Opus 4.6
Recent Highlights
- Claude Opus 4.6 expands enterprise reasoning with improved benchmarks and agent teams
- Claude Sonnet 4.6 brings near-flagship performance at mid-tier pricing
- Both models emphasize enhanced long-context reasoning and professional tool integration
Go Back
>Quit Program