Haiku 5.5 lands in Claude Code and Platform, running ~75% cheaper than Haiku 4.5
brada · x · 2026-10-08
Anthropic has made Haiku 5.5 available in the Claude Platform and Claude Code, with average running costs about 75% lower than Haiku 4.5. The company positions it as a subagent alongside Opus 5.5 or Sonnet 5.5 for high-volume, cost-sensitive tasks like summaries, compactions, and database queries.
Poster brada adds a hands-on data point: a team's design passed a validation run in 59 seconds for 15 cents using a multi-model setup, while Opus 5.5 alone took 2 minutes 4 seconds at 26 cents — with passing defined as surviving a 32 m drop straight, tilted, and slightly heavier.
Related event: Anthropic Launches Claude Haiku 5.5 at Roughly 75% Lower Cost(32 posts)→
More from Models
- repligate: Opus 3 and Opus 4 changes are being conflated — same root cause — repligate · 2026-10-08
- Chandra OCR team accuses interfaze-1-lite of copying their repo, worse scores — RexDouglass · 2026-10-08
- OpenAI rolls out GPT-6 with Intelligent UI that generates custom interfaces in ChatGPT — sama · 2026-10-08
- GPT-6 and Intelligent UI rolling out to all ChatGPT users, no GPT-6.1 yet — testingcatalog · 2026-10-08
- OpenAI rolls out new GPT-6 version to all ChatGPT users, citing model and infra upgrades to scale to 1.2B users — charliermarsh · 2026-10-08
- Anthropic confirms banked usage resets are coming to Claude — jeppa35 · 2026-10-08