Claude Sonnet 4.5

Claude Sonnet 4.5 is Anthropic's late-September 2025 upgrade to its coding-focused model line, arriving alongside Claude Code 2.0 with a native VS Code extension, terminal refresh, checkpoints, subagents, hooks, and background tasks. Anthropic positioned the release as a leap forward for software work, claiming major gains in reasoning, mathematics, and computer-use tasks compared with earlier Claude models. The model is intended for developers and teams building agentic coding workflows that need sustained reasoning across long sessions rather than just single-turn completions.

Within the Claude family, Sonnet 4.5 sits between the lighter Haiku 4.5 and the heavier Opus 4.5, aiming to deliver near-flagship coding and reasoning quality at a more accessible price point. Its practical fit is agentic engineering: multi-step refactors, codebase navigation, tool use, and structured planning where checkpoints and background tasks help preserve progress over long runs. For OpenCode Zen users selecting a balanced workhorse, the model is best suited when the job demands careful reasoning and reliable tool calling on non-trivial software problems, rather than simple or latency-critical chat.

OpenCode Zenclaude-sonnet-4-5claude-sonnet

Open provider page Provider docs

Quick Info

Provider: OpenCode Zen
Model key: claude-sonnet-4-5
Release date: Sep 29, 2025
Last updated: Sep 29, 2025
Knowledge cutoff: 2025-07-31
AI SDK package: @ai-sdk/anthropic
Input modalities
Output modalities
Capabilities

Cost

Input token cost: $3.00
Output token cost: $15.00

Limits

Output tokens: 64,000 tokens
Context window: 1,000,000 tokens

Transparent token rates

Compare Claude Sonnet 4.5 pricing

Rates are shown per one million tokens. Combined means one million input plus one million output tokens.

Browse this family

Latest news about Claude Sonnet 4.5

Venice AI

CoverageBenchmark

CodeRabbit's benchmark evaluation compared Claude Sonnet 4.5 against Sonnet 4 and Opus 4.1 across 25 difficult real-world pull requests with known critical bugs. Sonnet 4.5 achieved 41.5% "Important" comments versus 35.3% for Sonnet 4, demonstrating measurable progress in catching bugs and surfacing critical issues inc The evaluation found Sonnet 4.5 approaches Opus 4.1's coverage and precision at a fraction of the price, making it a pragmatic sweet spot for code review at scale. However, the model showed more hedging and self-questioning behavior, sometimes sounding like "a thoughtful colleague rather than a decisive reviewer." This

Venice AI

CoverageBenchmark

Cirra AI's technical analysis confirms Claude Sonnet 4.5 was released September 29, 2025, with a 200,000-token context window and up to 64K output tokens. The model uses a hybrid reasoning architecture with an "extended thinking" mode for multi-step tasks, enabling more sophisticated agentic workflows. Anthropic positi New API-side tools shipped alongside Sonnet 4.5 include context editing, memory, checkpoints, and VS Code integration to support long-running agentic workflows. The model demonstrated autonomous operation spanning 30+ hours on complex tasks. These capabilities reflect Anthropic's focus on making Sonnet 4.5 suitable for