Claude Haiku 5.5
Verdict: Anthropic's fastest small model, pitched for high-volume tasks and coding subagents.
Use this if you need cheap, fast calls for classification, support or subagents.
Live X Buzz Score
X 51/100 Steady updated 15 Oct, 15:13 HKT
Daily X posts, last 7 days (oldest → today): 4106 · 1600 · 1098 · 854 · 934 · 791 · 563. Score parts: volume 40/50, spike 0/25, engagement 10.6/25. Method
Why it's buzzing
@Xyz_Tox · 2026-10-11Claude Haiku 5.5 just dropped and it's dirt cheap 👀 Anthropic's new small model is also selectable for free users on https://t.co/uXYebdcpGn. - API: $0.10 per 1M input, $0.50 per 1M output (prompts up to 100K) - aroun
♥ 33 · ⟲ 3 · 💬 20 · 👁 145@simond_dev · 2026-10-13Claude Code 2.1.296 藏了个省钱开关： CLAUDE_CODE_WORKFLOW_SUBAGENT_MODEL 设一次，所有 workflow subagent 跑同一个模型，其他 subagent 各跑各的。 我的用法：设成 claude-haiku-5-5。 Haiku 5.5 现在 1M 上下文，短 prompt $0.10/$0.50——搜索、grep、读文件这种体力活，不值得上 Sonnet。
♥ 6 · ⟲ 0 · 💬 6 · 👁 103@quietoraclex · 2026-10-14holy sh*t, Claude Code can run on a model 40x cheaper than Opus now... Haiku 5.5 does the work. Opus 5.5 only wakes up for the hard calls: before a plan locks, when the same error hits twice, before it says "done" one
♥ 9 · ⟲ 0 · 💬 2 · 👁 310Alternatives
- Claude: Excellent for long documents, writing and careful coding help.
Discussion
Comments are coming with sign-in.