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 Sonnet 4.5AnthropicReleased September 29, 2025
Claude Sonnet 4.5
The 500k context window upgrade for Sonnet, combining high-speed execution with large codebase comprehension.
reasoning-llmProprietary API$3.00 / 1M tok (500k Context)Context: 500K (500,000 tokens)Arena ELO: 1435
Technical Specifications
Architecture Type
High-Efficiency 500k Transformer
Total Parameters
Frontier Scale
Context Window
500K (500,000 tokens)
Max Output Tokens
64K (64,000 tokens)
Knowledge Cutoff
August 2025
Supported Modalities
text, code, vision, reasoning
License & Access
Anthropic Commercial Terms
Benchmark Evaluations
Mmlu Pro
83.2
Gpqa Diamond
75.1
Swe Bench Verified
76.2
Chatbot Arena ELO
1435
Deep Architectural Overview
Claude Sonnet 4.5 expanded context retention to 500,000 tokens while maintaining low $3/$15 pricing. It introduced enhanced tool use definitions and native support for long-running autonomous development agents.
Strengths & Considerations
Core Strengths
- 500,000 token context window
- 76.2% SWE-bench score at standard $3/MTok price
- Prompt caching hit price of just $0.30 / 1M tokens
Known Limitations
- Active support through September 2026; succeeded by Claude Sonnet 5
Token & API Pricing
Input Tokens (1M)$3.00
Output Tokens (1M)$15.00
Cached Input (1M)$0.3000
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 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
The most capable autonomous intelligence system engineered by Anthropic, restricted to Project Glasswing participants.
1M ctx$10.00 / 1M tok (Glasswing Exclusive)
Anthropic
The premier enterprise flagship model for multi-hour autonomous coding, complex system engineering, and vision-heavy agent tasks.
1M ctx$5.00 / 1M tok (Opus 5 Flagship)