Claude Haiku 5.5 Ships Minutes Before GPT-6's ChatGPT Debut: 75% Cheaper, Tiered Pricing, First Haiku With Effort Controls
On Wednesday, October 7, 2026, Anthropic launched Claude Haiku 5.5 — the third Claude 5.5 model in a month (Opus 5.5 on Sept 22, Sonnet 5.5 on Sept 28), completing the family days before a planned IPO. Per Reuters: "Anthropic on Wednesday launched Claude Haiku 5.5, the third model in its Claude 5.5 family in the past month, expanding its artificial intelligence lineup ahead of a planned IPO."
The release is primarily a pricing weapon: $0.10 per million input / $0.50 per million output tokens for prompts under 100K tokens (cache reads $0.01), rising to $0.50/$2.50 above 100K — versus $1/$5 for Haiku 4.5, i.e. 10x cheaper in the tier where Anthropic says ~90% of Haiku requests live. Average running costs are ~75% below Haiku 4.5, and the headline price exactly matches GPT-6 Luna ($0.10/$0.50). Anthropic paired it with a 50% cut to Sonnet 5.5 cache-read pricing ($0.20 → $0.10/M, ~20% cheaper long-running agentic work) and new monthly API credits (Max 5x: $100, Max 20x: $200, Team: up to $500 pooled).
Capability claims carry a clean year-over-year anchor (Haiku 4.5 shipped Oct 15, 2025): Artificial Analysis scores it 43 on its Intelligence Index — up 26 points in one year — and it is the first Haiku with effort settings and adaptive thinking. Anthropic-reported jumps: OSWorld 2.1 offline 72.4% (vs 15.7% for Haiku 4.5), Terminal-Bench 4.0 39.2% (predecessor scored zero), Humanity's Last Exam 45.9% without tools / 57.4% with. Anthropic's positioning for it is the high-volume subagent: Mike Krieger — "Opus does the heavy thinking, Haiku does the high-volume work, and together you get more done for less."
Two caveats cut against the sticker price. First, token consumption: Neowin's writeup is titled "aggressive pricing, but it consumes far more tokens," and one developer found that at max effort it is "more token-hungry than opus 5.5 max." Second, safeguards: Reuters reports it is "Anthropic's first Haiku with built-in safeguards for a narrow set of high-risk cybersecurity requests" — per TestingCatalog, its cyber safeguards "allow more defensive work than Sonnet 5.5 but still block penetration testing," with biology protections matching the rest of the 5.5 line.
The timing was itself the message — Anthropic dropped Haiku 5.5 minutes before OpenAI's GPT-6-in-Chat announcement (see OpenAI's Post-DevDay Trust Problem: "28 Consecutive Days of Releases or Resets" — and the Sprint's Flagship Ship Is GPT-6 for All 1.2B Users), extending the price war tracked in OpenAI Answers Opus 5.5 Same-Day with GPT-6 Sol and Luna at 50% Cuts — the Price War Moves to Cost-per-Task and Claude Sonnet 5.5 Ships as a Deliberately Non-Frontier Release — and Beats Opus 5.5 on Terminal-Bench into the small-model tier. As one widely-shared reaction put it: "This is how you start a model war. Anthropic dropped Haiku 5.5 just minutes before OpenAI announced GPT-6 in Chat for everyone. Not hours. Minutes." Distribution was immediate: Claude Platform, AWS, Google Cloud, Azure, GitHub Copilot, and Cursor (where it costs 10x less than Haiku 4.5 on shorter requests).