New Left-Hand Navigation
Hatz moves its navigation into a sidebar on the left. Chat, Workshop, AI Phone Agent, Learn, and Workspace each open to list their pages, and the page you're on stays highlighted. Pages across Hatz also share consistent spacing and titles, with clearer empty states and tables that scroll cleanly.
LLMs
Claude Haiku 5.5
Claude Haiku 5.5 was added to the model selector for chats, apps, agents, and workflows.
- Credits:the same multiplier per message as Claude Haiku 4.5.
- Good for:classification, pulling details out of documents, chat, and tool use, with up to 1M tokens of context.
Gemini 4 Argon
Gemini 4 Argon was added to the model selector for chats, apps, agents, and workflows.
- How it compares:Google's benchmarks put it ahead of Claude Opus 5.5 and GPT-6 Astra on DeepSWE, a real-world software engineering benchmark.
- Good for:reasoning, coding, and long documents, with up to 1M tokens of context.
Model Catalog Maintenance
Older models have been retired so Hatz can focus on the most capable, most credit-efficient models.
- Nothing to update:any agent, workflow, or default that used a retired model now runs on its successor automatically.
- Same credits:each successor is priced the same as the model it replaces, so nobody pays more.
- Stay on Auto:Auto always routes to the current best model for the task.
Improved
Compact Tool Calls in Chat
When Hatz uses a tool in chat, each call shows as one line that says what it did, such as "Searched the web" or "Ran Python". Any member can click a line to see the code and output. A failed call shows in red and opens on its own.
Fixed
- Memory search:Hatz finds saved memories by meaning, even when a question uses different words.