Anthropic
Anthropic Claude model for long-running agentic coding and knowledge work.
1M ctxUSD 4 / 1M input tok; USD 20 / 1M output tok
Introducing Claude Opus 5AnthropicReleased July 24, 2026
Claude Opus 5
The premier enterprise flagship model for multi-hour autonomous coding, complex system engineering, and vision-heavy agent tasks.
reasoning-llmProprietary API$5.00 / 1M tok (Opus 5 Flagship)Context: 1M (1,000,000 tokens)Arena ELO: 1550
Technical Specifications
Architecture Type
Fifth-Generation Frontier Agentic Transformer
Total Parameters
Fifth-Gen Frontier Scale
Context Window
1M (1,000,000 tokens)
Max Output Tokens
128K (128,000 tokens)
Knowledge Cutoff
May 2026
Supported Modalities
text, code, vision, reasoning
License & Access
Anthropic Commercial Terms
Benchmark Evaluations
Math500
99.1
Mmlu Pro
92.4
Gpqa Diamond
87.1
Swe Bench Verified
86.8
Chatbot Arena ELO
1550
Deep Architectural Overview
Claude Opus 5 delivers a step-change improvement over Claude Opus 4.8, showing massive gains in test-time compute scaling and deep multi-step reasoning. Featuring 1M context, 128k output buffer, and mid-conversation tool dynamic updates, it powers the most sophisticated AI agents in the world.
Strengths & Considerations
Core Strengths
- World-class agentic coding (86.8% SWE-bench Verified)
- 1,000,000 token context window and 128k output buffer
- Supports dynamic mid-conversation tool modifications
- High-effort adaptive thinking on by default
Known Limitations
- Moderate latency compared to Sonnet 5, optimized for deep deliberation
Token & API Pricing
Input Tokens (1M)$5.00
Output Tokens (1M)$25.00
Cached Input (1M)$0.5000
Batch API Discount50% Off Standard
Pricing is verified directly against Anthropic's developer documentation and API rate sheets.
Similar & Alternative Models
Explore other frontier models from Anthropic and comparable reasoning engines.
Anthropic
The most capable autonomous intelligence system engineered by Anthropic, restricted to Project Glasswing participants.
1M ctx$10.00 / 1M tok (Glasswing Exclusive)
Anthropic
The ultimate model for demanding reasoning and long-horizon agentic work, featuring 0.025x cache reads ($0.25/MTok).
1M ctx$10.00 / 1M tok (0.025x Cache Reads)
Anthropic
Anthropic’s fifth-generation workhorse, combining 1M context with adaptive thinking on by default at a lower $2/$10 price point.
1M ctx$2.00 / 1M tok (1M Context Sonnet 5)