Home
Anthropic Launches Claude Haiku 5.5, Cutting Small Model Costs by 75% and Letting Users Adjust Inference Intensity

On October 8th, Anthropic introduced its most efficient and capable model to date: Claude Haiku 5.5. Positioned as the fastest and most capable small model in Anthropic’s lineup, it is engineered for high-frequency, cost-sensitive applications such as article summarization, content categorization, database queries, customer support responses, browser automation, and sub-tasks delegated by larger AI agents.
Haiku 5.5 adopts a highly competitive strategy, reducing average operating costs by approximately 75% compared to its predecessor, Haiku 4.5. For requests under 100,000 tokens, input and output costs are just $0.1 and $0.5 per million tokens, respectively—significantly lowering expenses for lightweight tasks. Notably, the Haiku series now offers adjustable reasoning intensity, allowing users to control the depth of thinking based on task complexity. This feature enables the affordable model to perform deeper reasoning when necessary, without incurring full computational costs for every request.
Anthropic also adjusted pricing for its neighbor model, Sonnet 5.5, halving the cache reading price to $0.1 per million tokens. According to the company, this change reduces operational costs for most agent tasks by about 20%. Cache reading, frequently used when agents repeatedly process long contexts, is now more affordable, effectively lowering the cost of running autonomous agents in the background throughout the day.
Related article
Otter.ai Secures $320 Million Growth Funding From General Catalyst for Meeting Tool
On Tuesday, Otter.ai, a meeting note-taking platform, announced a $320 million investment from General Catalyst’s Customer Value Fund (CVF). This fund supports startups without requiring equity dilution, instead seeking returns based on marketing per
Claude Unveils First Full Ultraviolet Sky Map, Charting 119 Million Stars
With the powerful assistance of Claude, Anthropic’s large language model, astrophysicists have successfully generated the first complete ultraviolet all-sky map in history. This vast star chart, featuring 119 million stars, was largely completed by C
How AWS and NVIDIA Are Powering Agentic and Physical AI
NVIDIA CEO Jensen Huang (Credit: Tomohiro Ohsumi/Getty Images)Jensen Huang’s NVIDIA and AWS scale their global partnership, deploying 2M GPUs to power next-generation agentic and physical AI enterprise infrastructureAWS and NVIDIA have announced a ma
Related Special Topic Recommendations
Comments (0)
0/500

On October 8th, Anthropic introduced its most efficient and capable model to date: Claude Haiku 5.5. Positioned as the fastest and most capable small model in Anthropic’s lineup, it is engineered for high-frequency, cost-sensitive applications such as article summarization, content categorization, database queries, customer support responses, browser automation, and sub-tasks delegated by larger AI agents.
Haiku 5.5 adopts a highly competitive strategy, reducing average operating costs by approximately 75% compared to its predecessor, Haiku 4.5. For requests under 100,000 tokens, input and output costs are just $0.1 and $0.5 per million tokens, respectively—significantly lowering expenses for lightweight tasks. Notably, the Haiku series now offers adjustable reasoning intensity, allowing users to control the depth of thinking based on task complexity. This feature enables the affordable model to perform deeper reasoning when necessary, without incurring full computational costs for every request.
Anthropic also adjusted pricing for its neighbor model, Sonnet 5.5, halving the cache reading price to $0.1 per million tokens. According to the company, this change reduces operational costs for most agent tasks by about 20%. Cache reading, frequently used when agents repeatedly process long contexts, is now more affordable, effectively lowering the cost of running autonomous agents in the background throughout the day.
Otter.ai Secures $320 Million Growth Funding From General Catalyst for Meeting Tool
On Tuesday, Otter.ai, a meeting note-taking platform, announced a $320 million investment from General Catalyst’s Customer Value Fund (CVF). This fund supports startups without requiring equity dilution, instead seeking returns based on marketing per
Claude Unveils First Full Ultraviolet Sky Map, Charting 119 Million Stars
With the powerful assistance of Claude, Anthropic’s large language model, astrophysicists have successfully generated the first complete ultraviolet all-sky map in history. This vast star chart, featuring 119 million stars, was largely completed by C
How AWS and NVIDIA Are Powering Agentic and Physical AI
NVIDIA CEO Jensen Huang (Credit: Tomohiro Ohsumi/Getty Images)Jensen Huang’s NVIDIA and AWS scale their global partnership, deploying 2M GPUs to power next-generation agentic and physical AI enterprise infrastructureAWS and NVIDIA have announced a ma











