Claude Haiku 5.5: the newest model is live
A new model has arrived on NeuralSpace — Claude Haiku 5.5. It is already available in the chat, the Code section and the public API: the fastest and most cost-effective model of the Haiku line with a 1M token context. The model page is at /chat/claude-haiku-5-5.
What Claude Haiku 5.5 can do
- The fastest and most cost-effective. The newest Haiku answers almost instantly and costs less than the senior models — a good fit for large volumes of everyday tasks.
- 1M token context. Attach a large document or keep a very long conversation going — the model keeps track of it.
- Web search. Like the neighbouring chat models: you can ask it to look up current information.
- Coding. It reads existing code, plans changes and applies them step by step, explaining the result.
Pricing
Cost is based on the tokens actually used and is charged after the answer — the minimum balance threshold is shown before you send, just like for the other models in the section. Haiku 5.5 has a long-context tier: up to 100,000 input tokens the base rate applies, and above that threshold a higher rate is used. Current prices are on the API page.
How to try it
- Open the chat and pick Claude Haiku 5.5 from the model list — the model page is also available at /chat/claude-haiku-5-5.
- Ask as usual: the model streams its answer and supports web search.
- For bigger projects open the Code section, where the model works as an agent: it creates a project, writes and edits files and runs long iterations.
When to choose Haiku 5.5
Haiku 5.5 is an everyday working model: fast answers, text and document analysis, day-to-day coding, high-volume repetitive tasks. When a task calls for maximum depth, the senior models such as Claude Sonnet 5.5 and Claude Opus 5.5 are there for that.
Frequently asked questions
How is Haiku 5.5 different from Haiku 4.5? It is the newest generation of the line: faster and better answers, with a 1M token context.
Where is it available? In the chat, the Code section, and through the public API with a key — see the model list and prices on the API page.
Does it have web search? Yes, like the neighbouring chat models: you can ask it to look up current information.
How much does it cost? You pay for the tokens actually used, the price is shown before you send, and requests above 100,000 input tokens use the higher long-context tier rate.