IT之家 October 8 news — Anthropic published a blog post yesterday (October 7) announcing the launch of the Claude Haiku 5.5 model, calling it the fastest, cheapest, and most capable small model in the Haiku series to date.
On pricing, IT之家 cites the blog post and provides the price information below:
| Price per 1 million Tokens | Haiku 5.5 (prompts up to / over 100k) | Haiku 4.5 | Sonnet 5.5 |
|---|---|---|---|
| Cache read | 0.01 USD (IT之家 note: approximately 0.067 RMB at current exchange rate) / 0.05 USD (approximately 0.34 RMB at current exchange rate) | 0.10 USD (approximately 0.67 RMB at current exchange rate) | 0.10 USD (approximately 0.67 RMB at current exchange rate) |
| Cache write | 0.125 USD (approximately 0.84 RMB at current exchange rate) / 0.625 USD (approximately 4.2 RMB at current exchange rate) | 1.25 USD (approximately 8.4 RMB at current exchange rate) | 2.50 USD (approximately 16.8 RMB at current exchange rate) |
| Input Tokens | 0.10 USD (approximately 0.67 RMB at current exchange rate) / 0.50 USD (approximately 3.4 RMB at current exchange rate) | 1.00 USD (approximately 6.7 RMB at current exchange rate) | 2.00 USD (approximately 13.4 RMB at current exchange rate) |
| Output Tokens | 0.50 USD (approximately 3.4 RMB at current exchange rate) / 2.50 USD (approximately 16.8 RMB at current exchange rate) | 5.00 USD (approximately 33.6 RMB at current exchange rate) | 10.00 USD (approximately 67.1 RMB at current exchange rate) |
Haiku 5.5's running cost is on average about 75% lower than Haiku 4.5. For requests of up to 100,000 Tokens, the input price is 0.10 USD (approximately 0.67 RMB at current exchange rate) per million Tokens, and the output price is 0.50 USD (approximately 3.4 RMB at current exchange rate). Beyond 100,000 Tokens, the input and output prices are 0.50 USD (approximately 3.4 RMB at current exchange rate) and 2.50 USD (approximately 16.8 RMB at current exchange rate) respectively.
Anthropic has simultaneously lowered the cache read price of Sonnet 5.5. The price has dropped from 0.20 USD (approximately 1.3 RMB at current exchange rate) to 0.10 USD (approximately 0.67 RMB at current exchange rate) per million Tokens. Anthropic said this adjustment can reduce Sonnet 5.5's running cost in most agentic tasks by about 20%.
In knowledge work tests, Haiku 5.5 scored 1620 on GDPval-AA v2.1 and 1578 on AA-Briefcase v1.1. It scored 72.4% on the OSWorld 2.1 offline subset and 39.2% on Terminal-Bench 4.0.
Haiku 5.5 is the first Haiku model to support adjustable performance settings, allowing users to choose a configuration for cost and intelligence level. Anthropic also positions it as a subagent model for Opus 5.5 and Sonnet 5.5 when they execute coding tasks.
