Claude Haiku 5.5

by Anthropic

Claude Haiku 5.5 is Anthropic's most capable, cost-efficient, and fastest small model, released October 7, 2026. It features a 1-million-token context window, 128K maximum output, and a two-tier pricing structure based on prompt length. The first Haiku model with adjustable thinking effort, it is designed for high-volume, latency-sensitive tasks such as classification, data extraction, and subagent workloads. Available via Claude API, Amazon Bedrock, Google Cloud, and Microsoft Foundry.

Large Language Models
Proprietary

Primary Use Cases

  • High-volume classification tasks
  • Data extraction pipelines
  • Subagent workloads in multi-agent systems
  • Latency-sensitive applications
  • Cost-optimized AI workflows

✅ Strengths

  • ~75% cheaper than Claude Haiku 4.5 on average
  • First Haiku with thinking/reasoning capability
  • Available across major cloud platforms (Bedrock, Google Cloud, Microsoft Foundry)
  • Optimized for speed and cost efficiency

⚠️ Limitations

  • Does not support manual extended thinking via budget_tokens
  • Non-default temperature/top_p/top_k settings return errors
  • No assistant message prefill support
  • New tokenizer may produce ~30% higher token counts vs. Haiku 4.5
  • Conversations must be append-only

Technical Specifications

Developer
Anthropic
Category
Large Language Models
Type
🧠 Foundation Model
License
UNKNOWN
Version
Haiku 5.5

Trust & Privacy

Health Status
🟢 Active
Trains on User Data
🔄 Opt-out available
Certifications
None listed

Ecosystem & Integrations

API Access
✅ Available
Chrome Extension
❌ No
Mobile App
❌ No
Pricing Model
💰 Pay Per Use

Pricing Breakdown

Standard (up to 100K tokens)

For prompts up to 100,000 tokens

Input$0.1/M tokens
Output$0.5/M tokens
0
Extended (over 100K tokens)

For prompts exceeding 100,000 tokens

Input$0.5/M tokens
Output$2.5/M tokens
0