Anthropic released the Claude Haiku 5.5 model on October 7, 2026. It is designed for high-volume, cost-sensitive tasks such as database queries, automation, and customer support. The model features a 1M-token context window, a significant increase from the 200k window in the previous version, and supports browser tools and subagent delegation.
In the Agent Arena leaderboard, Claude Haiku 5.5 (High) holds the 27th position as of mid-October 2026. It also ranks 30th in the WebDev category of Code Arena. The model achieved 1620 points on the GDPval-AA v2.1 benchmark and a 72.4% success rate on OSWorld 2.1 offline, demonstrating its capability in operating computer environments.
The API pricing for the model starts at $0.10 per million input tokens and $0.50 per million output tokens for prompts under 100k tokens. Anthropic notes that average costs are approximately 75% lower than those of the predecessor model. The pricing structure is tiered based on prompt length.
Claude Haiku 5.5 is available through the Anthropic API, Claude.ai, Amazon Bedrock, Google Cloud, and Microsoft Foundry. It serves as a lighter alternative to the heavier Opus 5.5 and Sonnet 5.5 models, allowing teams to delegate repetitive tasks while reserving more complex reasoning for the larger models.
