Teamday

Chat & Reasoning Model · Anthropic

Claude Haiku 5.5

Claude Haiku 5.5 is Anthropic's newest Haiku, released October 7, 2026, for high-volume, latency-sensitive work such as classification, extraction, routing, and sub-agent tasks. It succeeds Claude Haiku 4.5 with a 1M-token context window (up from 200K), up to 128K output tokens, and adaptive thinking with selectable effort. On prompts up to 100K tokens it costs a tenth of Haiku 4.5; longer prompts pay a higher rate.

Use Claude Haiku 5.5 in Teamday →All models

At a glance (verified 2026-10-07)

Vendor
Anthropic
Type
Chat & Reasoning Model
Context window
1M tokens
Pricing
Prompts up to 100K tokens: $0.10 in / $0.01 cache hits / $0.50 out per 1M tokens; cache writes $0.125 5m / $0.20 1h. Prompts over 100K tokens: $0.50 in / $0.05 cache hits / $2.50 out; cache writes $0.625 5m / $1 1h.

Best for

Real-time chatHigh-volume tasksClassification and extractionSub-agent workCost control

How to use Claude Haiku 5.5 in Teamday

Select Claude Haiku 5.5 in Teamday's chat model picker on the Claude Code harness. Agents and chats previously set to Haiku 4.5 run on it automatically.

Harness: claude-code →James →Sarah →

Frequently asked questions

How much does Claude Haiku 5.5 cost per million tokens?

It depends on prompt length. For prompts up to 100K tokens: $0.10 per 1M input tokens, $0.01 for cache hits, and $0.50 per 1M output tokens. For prompts over 100K tokens: $0.50 input, $0.05 cache hits, and $2.50 output. Cache writes cost 1.25x input (5-minute) or 2x input (1-hour).

What changed from Claude Haiku 4.5?

A 1M-token context window instead of 200K, 128K output tokens instead of 64K, and reasoning effort from low to max (Teamday starts it at medium, Anthropic's default). It uses Anthropic's newer tokenizer, which produces about 30% more tokens for the same text.

When should I pick Claude Haiku 5.5 over Sonnet 5.5?

Pick Haiku 5.5 for quick answers and high-volume, repeatable work where speed and cost matter most. For everyday coding and multi-step agent work, Anthropic positions Sonnet 5.5 as the better balance of speed and intelligence.

Is Claude Haiku 5.5 available in Teamday?

Yes — it is in Teamday's chat model picker on the Claude Code harness, with reasoning effort selectable from low to max.

Related

Claude Haiku 5.5 is one of the models Teamday's AI employees already run — no API keys, no infrastructure.

Hire your first AI employee →