Breaking

Anthropic Releases Claude Haiku 5.5 With Lower Pricing and 1M-Token Context

October 7, 2026 at 19:49 EDT

Foundation Models
AI Agents
Software Dev & Coding

Anthropic released Claude Haiku 5.5 on October 7, positioning its smallest Claude tier as a faster and cheaper worker model for high-volume AI applications. The launch gives developers a new claude-haiku-5-5 model with a 1M-token context window, up to 128K output tokens, and pricing that Anthropic says is about 75% lower to run on average than Claude Haiku 4.5.

October 7, 2026 · Anthropic

Claude Haiku 5.5 moves into the cheap-agent layer

The new small Claude model targets high-volume work with lower token prices, a 1M-token context window, and stronger results on computer-use and agent benchmarks.

Average run cost vs. Haiku 4.5

100%

Haiku 4.5 baseline

25%

Haiku 5.5 average

Average run cost is 75% lower than the Haiku 4.5 baseline.

context tokens

128K

max output tokens

$0.10

input per 1M tokens

90%

prior requests under 100K

Computer-use benchmark jump

15.7%

Haiku 4.5

72.4%

Haiku 5.5

83.9%

Sonnet 5.5

OSWorld 2.1 offline subset shows Haiku narrowing the gap on computer-use tasks while Sonnet remains ahead.

Where the model fits

Route

→

Extract

→

Summarize

→

Hand back

Haiku 5.5 is designed to handle bounded subagent work around larger Claude models.

Positive signal

Customer tests reported lower latency, higher scores, and practical uses in CRM audits, document Q&A, enterprise content, and coding sidekick workflows.

Limits to watch

Long prompts move to a higher price tier, the tokenizer counts more tokens than Haiku 4.5, and Sonnet or Opus remain better suited for complex agentic coding.

Continue reading

The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.