Artificial Intelligence
Anthropic launches Claude Haiku 5.5
Anthropic introduced Claude Haiku 5.5, said it was its cheapest and fastest small model, and announced lower Sonnet 5.5 cache-read prices plus new monthly API credits.
- Anthropic said Claude Haiku 5.5 was its cheapest, fastest and most capable small model.
- The company said Haiku 5.5 cost about 75% less to run on average than Haiku 4.5.
- Anthropic said it was halving Claude Sonnet 5.5 cache-read prices.
- The company said Haiku 5.5 was available on Amazon Web Services, Google Cloud and Microsoft Azure, and on the Claude Platform as claude-haiku-5-5.
- Anthropic said Haiku 5.5 had tighter cybersecurity safeguards than Haiku 4.5, but looser ones than some recent models.
Anthropic introduced Claude Haiku 5.5 and described it as its lowest-cost, quickest and most capable small model to date. It said the model was aimed at repetitive, high-volume work such as summaries, compaction, database queries and classification, and could also serve as a subagent for coding tasks.
The company said Haiku 5.5 cost around 75% less to run on average than Haiku 4.5. It also said the model had an adjustable effort setting, allowing users to trade cost against capability.
Anthropic said early customers reported results in line with its claims about performance and cost. It also said Sonnet 5.5 cache reads were being cut by half, which it said would lower the cost of most agentic work by about 20%.
On safety, Anthropic said Haiku 5.5 had stronger alignment results than Haiku 4.5 and more restrictive cybersecurity safeguards than Haiku 4.5, though less restrictive than some of its newer models. It said the safeguards allowed defensive work but blocked penetration testing and similar techniques.
Named in this story
Companies
- Anthropic
- released Claude Haiku 5.5 and changed pricing across its model range
- Amazon Web Services
- carried the model at launch
- Google Cloud
- carried the model at launch
Products and systems
- Claude Haiku 5.5
- the new small model Anthropic launched
- Claude Haiku 4.5
- the earlier model used for price and safety comparisons
- Claude Sonnet 5.5
- had its cache-read pricing cut and was compared with Haiku 5.5
- Claude Opus 5.5
- was used in capability and safety comparisons
- Claude Platform
- the platform where developers could start using the new model
- Microsoft Azure
- carried the model at launch
How the source tells it
The article read as a product launch note, with repeated self-promotional claims, benchmark-style assurances and a separate safety section meant to temper the sales pitch.
- hype Superlative claims were used to present the model as the best in its class without outside confirmation.
- vendor boosterism The company repeated first-person praise of its own model, pricing and customer take-up throughout the release.
- reassurance A safety section stressed tighter safeguards and limits on misuse to offset the launch messaging.