26.03 Release Notes
August 4th, 2026
š¤ Better Anthropic Model Support
Working with Claude in Max just got noticeably smoother. This release brings a set of improvements that make Anthropic models more reliable end to end:
- Clean rendering in chat. Claude's responses now display exactly as intendedāno more blank rows or garbled entries in your conversation history.
- Full fidelity in the Playground. Original Claude answers load properly, so you can review, compare, and refine with confidence.
- Graceful handling when things get busy. If Anthropic's service is overloaded, rate-limited, or slow to respond, Max now retries intelligently behind the scenes and surfaces a clear, plain-English message instead of a technical error.
š Dedicated OIDC Group Mappings
OIDC group-to-permission mappings now have a home of their own, managed separately from SAML with a dedicated tab in the Admin UI. Permissions are applied automatically the moment a user's OIDC groups matchāno extra toggle to remember, no manual follow-up. Cleaner setup, clearer intent, fewer access surprises.
Additional Note for Self-Hosted Customers
This release does not introduce additional upgrade requirements beyond standard processes. As always, we recommend running backups before upgrades.
To download Max from our container registry use the tag: v26.03.0
Tested Models
AnswerRocket SDK version: 0.2.107
OpenAI:
- Chat: gpt-5.1-2025-11-13
- Narrative: gpt-5.1-2025-11-13
- Evaluations: gpt-5.1-2025-11-13
- SQL: o4-mini-2025-04-16
- Researcher: gpt-5-mini-2025-08-07
- Embedding: text-embedding-3-small
Gemini:
- Chat: gemini-2.5-pro
- Narrative: gemini-2.5-pro
- Evaluations: gemini-2.5-pro
- Embedding: models/embedding-001
Anthropic:
- Chat: claude-opus-4-8
- Narrative: claude-opus-4-8
- Evaluations: claude-opus-4-8
- Embedding: Not Applicable