Claude Sonnet 4.5

The llm-stats.com page tracks Claude Sonnet 4.5 (identifier claude-sonnet-4-5-20250929) across its composite leaderboard and conversation-depth metrics, ranking it among 117 tracked models. Benchmarks cited include MMMLU (multilingual Q&A, rank 13 with a score of 0.89) and AIME 2025 (all 30 problems from the 2025 Ameri Cost-efficiency data lists Claude Sonnet 4.5 at a blended price of $3.57 per million tokens against an LLM Stats Score of 32.1, placing it alongside other frontier entries such as Claude Opus 5.5 ($4.76) and DeepSeek-V4.1-Flash ($0.24). A Quality Tracker line records 746 votes over a 7-day window with a stable +2.61σ s

Cirra AI's October 13, 2025 technical analysis confirms Claude Sonnet 4.5 targets coding, agentic workflows, and complex computer use, calling it a major new LLM in that space. The model is described as a hybrid reasoning LLM with an expanded 200,000-token context window and up to 64K output tokens, supporting multi-st The analysis highlights new tooling that supports long-running agents, including context editing, a memory tool, checkpoints, and VS Code integration in the Claude API. These additions are presented as enabling agents to operate across many turns and even beyond the context window, which Cirra frames as the strategic d

Anthropic announced Claude Sonnet 4.5 on September 29, 2025, positioning it as its strongest model for coding, complex agents, and computer use, with substantial gains in reasoning and math. On the SWE-bench Verified evaluation it is state-of-the-art, and the company reports it maintained focus for over 30 hours on mul The release shipped with major product upgrades: Claude Code gained checkpoints, a refreshed terminal interface, and a native VS Code extension; the Claude API received a context-editing feature and memory tool enabling longer-running agents; and the Claude apps gained inline code execution and file creation (spreadshe