Anthropic Launches Claude Haiku 5.5 and Cuts Sonnet 5.5 Cache Pricing
10/08/2026 — 10/11, 05:19·11 sources (all)·14 reports (all)·In the 2026-10-09 briefing·In the 2026-10-08 briefing
Key takeaways
- Anthropic has released Claude Haiku 5.5, its new lightweight model, now available on the Claude API, Amazon Bedrock, Google Cloud and Micros…
Background & analysis
On October 7, 2026, Anthropic released Claude Haiku 5.5, a new small model, alongside a cut to the cache-read price of Claude Sonnet 5.5. Haiku 5.5 is available on the Claude Platform, Amazon Bedrock, Google Cloud, Microsoft Foundry and Claude Platform on AWS under the model ID claude-haiku-5-5, or anthropic.claude-haiku-5-5 on Bedrock. Anthropic describes it as the fastest and most efficient model in the Claude 5.5 family, built for subagents and high-volume, cost-sensitive work such as summarization, data cleanup, database queries, classification and live customer support. Several reports put its average running cost at roughly 75% below Haiku 4.5.
Pricing depends on prompt length. Requests up to 100,000 tokens cost $0.10 per million input tokens and $0.50 per million output tokens; beyond that threshold, the rates rise to $0.50 and $2.50. Against Haiku 4.5's $1 and $5, that is a 90% cut within the tier and a 50% cut above it. One report calculated that a 100,000-token prompt with 2,000 output tokens costs $0.011, while adding a single token to the prompt raises the same request to $0.055. Another noted that the new tokenizer consumes about 30% more tokens for the same text, eating into some of the savings.
On performance, The Decoder reported that Haiku 5.5 jumped from 15.7% to 72.4% on the OSWorld computer-use test. Zhidx said it scored 43 on Artificial Analysis's AI analysis index, slightly ahead of GLM-5.3 Flash at 42, Gemini 3.8 Flash at 41 and GPT-6 Luna at 38. Open Source China reported that it is the first Haiku model with an adjustable effort setting, defaulting to medium in the API, and that its agent capability rose from 0 to 39%. Leiphone noted it still trails on complex coding tasks.
The same day, Anthropic lowered Sonnet 5.5 cache reads from $0.20 to $0.10 per million tokens, effective immediately, and introduced monthly API credits for Max and Team subscribers: $100 for Max 5x, $200 for Max 20x and up to $500 shared across Team users. All of these changes are now in effect, with availability and tiered pricing as published by Anthropic.
AI-generated from 14 reports · updated 34 days ago
Latest turnAnthropic released Claude Haiku 5.5, priced at $0.10 per million input tokens and $0.50 per million output tokens for prompts under 100,000 tokens — 90% below Haiku 4.5. Past that threshold the rates rise to $0.50 and $2.50, a 50% cut. Its new tokenizer also consumes about 30% more tokens for the same text, and it is the first Haiku with an effort setting, defaulting to medium.
Related AI tools
- Heat index
- 25
- All sources
- 11
- All reports
- 14
- First seen
- 45 days ago
Reports on this story headlines open the original
Anthropic released Claude Haiku 5.5, priced at $0.10 per million input tokens and $0.50 per million output tokens for prompts under 100,000 tokens — 90% below Haiku 4.5. Past that threshold the rates rise to $0.50 and $2.50, a 50% cut. Its new tokenizer also consumes about 30% more tokens for the same text, and it is the first Haiku with an effort setting, defaulting to medium.
DEV Community · AIAI score 60
Anthropic has released Claude Haiku 5.5, which dynamically allocates reasoning effort per task and shows clear gains in agentic operations while still trailing on complex coding. According to Leiphone, both input and output prices rise fivefold beyond the 100k-token mark.
雷峰网AI score 55Claude Haiku 5.5 is about 75% cheaper than its predecessor, and a zero-dependency Python routing calculator shows how to send lightweight work to Haiku while reserving Sonnet for heavier tasks. The author estimates roughly 60% savings across 10,000 daily calls and warns that the new tokenizer quietly raises effective prices.
掘金AI score 61Anthropic released Claude Haiku 5.5 on October 7, according to OSChina, calling it the first Haiku model with an adjustable effort setting so users can trade speed against intelligence. Pricing drops to about a quarter of Haiku 4.5: $0.10 per million input tokens and $0.50 per million output tokens for prompts under 100k tokens. The report says its agent capability rose from 0 to 39%.
开源中国AI score 65Anthropic has released Claude Haiku 5.5, aimed at high-volume projects where speed and low cost matter most, according to Golem.de. The company positions the new Haiku release as both cheaper and faster.
Golem.deAI score 53Anthropic has released Claude Haiku 5.5, the third model in the Claude 5.5 family, aimed at high-concurrency, low-latency and cost-sensitive workloads. API pricing drops sharply, putting it directly into the low-cost model race, though the new tokenizer consumes more tokens.
cnBetaAI score 59Anthropic has released Claude Haiku 5.5, its cheapest and fastest mini model yet, with average running costs roughly 75% lower than Claude Haiku 4.5 and pricing down 90%. The company says it targets high-throughput, cost-sensitive work such as summarization, data wrangling, database queries and real-time support, and can run as a coding sub-agent alongside Sonnet 5.5. It scores 43 on Artificial Analysis's AI analysis index, edging out GLM-5.3 Flash and Gemini 3.8 Flash.
智东西AI score 72Anthropic released Claude Haiku 5.5 on October 7, and it is now available on the Claude API, Amazon Bedrock, Google Cloud and Microsoft Foundry at $0.10 per million input tokens, down from $1 for Haiku 4.5. The cheap rate applies only to prompts up to 100,000 tokens, though: add one more token and the same request jumps from $0.011 to $0.055.
DEV Community · AIAI score 66Anthropic lowered Claude Sonnet 5.5 cache read prices by 50% to $0.10 per million tokens and introduced monthly platform API credits for Max and Team subscribers alongside the launch of Claude Haiku 5.5.
Unite.AIAI score 70Anthropic has reduced the cache read price for Sonnet 5.5 from $0.20 to $0.10 per million tokens while introducing monthly API credits ranging from $100 to $500 for Max and Team users.
TechmemeAI score 56Anthropic announced the availability of Claude Haiku 5.5 on Amazon Bedrock and Claude Platform. Designed for subagents and high-volume tasks, it is touted as the fastest and most efficient model in the family, costing around 75% less than Claude Haiku 4.5.
AWS Machine Learning BlogFirst-partyAI score 72Anthropic has released Claude Haiku 5.5, its fastest and most affordable small model to date, alongside price cuts for Sonnet 5.5. Haiku 5.5 targets high-volume, cost-sensitive tasks with price reductions of up to 90% for prompts under 100k tokens, though an updated tokenizer may increase overall token consumption.
The DecoderAI score 74According to The New Stack, Anthropic launched Claude Haiku 5.5, the first update to its lightweight model in nearly a year. The model is priced at $0.10 per 1M input and $0.50 per 1M output tokens for requests up to 100K tokens, with higher rates applying beyond that threshold.
TechmemeAI score 79Anthropic has released Claude Haiku 5.5, its newest small model available immediately across major platforms at approximately 75% lower running costs than its predecessor.
Unite.AIAI score 76
Other stories people are talking about
- Heat index 1004Anthropic Mathematician Says Claude Formalized the Generalized Riemann Hypothesis21 sources
- Heat index 135SurgeAnthropic Details Claude Agent Memory System and Dreaming Mechanism2 sources
- Heat index 593Anthropic IPO total raise reaches $93.5 billion after greenshoe exercise19 sources
- Heat index 37Developer uses hooks to push Claude Code alerts to iPhone1 sources
- Heat index 67Developer shares five Claude Code hooks that block recurring mistakes1 sources
How is heat calculated?About the methodHide
Heat counts how many independent sources covered a story in the last 48 hours: one source counts once no matter how many posts it published, decaying with a 24-hour half-life. What ranks first is what many people are talking about.
This page aggregates public feeds. Headlines and summaries are machine-organized and remain the property of the original authors; verify important facts at the source. If you believe a headline or summary infringes your rights, email the contact address in the footer with the page URL and basis of your claim, and we will remove or replace it after review.
- Surge
- Discussion rising fast
- New
- First report within 6 hours
- Rising
- Still gathering discussion

