Skip to content

Claude Haiku 4.5

claude-haiku-4-5-20251001
Anthropic · reasoning · api · seen 13m ago
GA Alert me on changes
Context
200K
Max output
64K
Input $/1M
$1.00
Output $/1M
$5.00
Modalities
text · image · document
Released
15 Oct 2025
AI summary
● machine-written

Claude Haiku 4.5: Anthropic's fastest model with extended thinking

Claude Haiku 4.5 is Anthropic's first Haiku-tier model to include extended thinking, computer use, and context awareness capabilities. It delivers performance comparable to Claude Sonnet 4 at approximately one-third the cost, with a 200K token context window and up to 64K output tokens. The model processes text and images and is optimized for high-volume, cost-sensitive deployments.

What's new
  • Extended thinking enables reasoning before response generation
  • Computer use capabilities for screenshots, clicks, and form filling
  • Context awareness allows model to track its own token consumption
  • Achieves 73.3% on SWE-bench Verified, within 5 percentage points of Sonnet 4.5
  • 50.7% score on OSWorld for computer use benchmarks
Best for
Free-tier products where cost per request is criticalParallelized sub-agents and high-volume pipelinesMulti-agent orchestration and scaled deploymentsReal-time applications requiring fastest latency
Sources

Source: https://docs.claude.com/en/docs/about-claude/models/overview