Move to Claude 5.5 models

[hanna-paasivirta]

Short Description

Move to the Claude 5.5 models.

Closes OFN-4756

Implementation Details

Into release/3.5.0, which already has the bumps from #716.

Models

• IDs: claude-opus-5-5, claude-sonnet-5-5, claude-haiku-5-5 in models.py. Routes keep their tier.
• Effort: unchanged. Router and docs retrieval tested at low against the default with the same decisions.

Fixes the new models need

• Text by block type: the router, docs retrieval, vocab_mapper and the test judge read content[0], which is now a thinking block.
• max_tokens raised instead of lowering effort, since thinking counts toward it: router 4096, docs retrieval and vocab_mapper 4096, planner, job_chat, workflow_chat and doc_agent_chat 32768. workflow_chat and doc_agent_chat get the explicit timeout the SDK requires above ~21k.
• Refusals: a decline is a 200 with stop_reason: "refusal". The planner, job_chat and workflow_chat now raise 422 MODEL_REFUSAL with the category instead of an empty-output error, and workflow_chat no longer retries it.
• Planner narration: notes between tool calls now come back as thinking blocks, so the planner sets display: "updates" and streams them as narration.

Behaviour changes still to tune

• Router: "what does this step do" on a step page now often goes to the planner instead of job_chat.
• Planner: narrates less before tool calls.
• job_chat: 2.5–4× more output tokens and about twice as slow.

AI Usage

• Yes, I have used AI
• No, I have not used AI

You can read more details in our
Responsible AI Policy