StockDuty ← Dashboard Oct 07, 2026 11:41 PM ET
SignalTech/AI

Anthropic cut Haiku pricing 90% — $0.10/$0.50 per million tokens vs $1/$5 for Haiku 4.5 — and it lands exactly on OpenAI's GPT-6 Luna rate. Sonnet 5.5 cache-read gets halved to $0.10/million, plus mon

Anthropic launches cheapest, fastest Claude model, signaling aggressive AI price-performance competition.

What happened: Anthropic released Claude Haiku 5.5, calling it its cheapest, fastest and most capable small model, available on the Claude website, AWS, Google Cloud and Microsoft Azure as claude-haiku-5-5. Key numbers: $0.10 per million input tokens and $0.50 per million output tokens for prompts up to 100,000 tokens, down Why it matters: The pricing matches OpenAI's GPT-6 Luna and undercuts Haiku 4.5 by 90%, with Anthropic putting average savings near 75%. Benchmarks place Haiku 5.5 ahead of GPT-6 Luna on OSWorld 2.1 and Terminal-Bench 4.0 but behind Sonnet 5.5's 70.6% on the coding test.
Anthropic released Claude Haiku 5.5, calling it its cheapest, fastest and most capable small model, available on the Claude website, AWS, Google Cloud and Microsoft Azure as claude-haiku-5-5.
The pricing matches OpenAI's GPT-6 Luna and undercuts Haiku 4.5 by 90%, with Anthropic putting average savings near 75%. Benchmarks place Haiku 5.5 ahead of GPT-6 Luna on OSWorld 2.1 and Terminal-Bench 4.0 but behind Sonnet 5.5's 70.6% on the coding test.
Sources