Skip to content

feat(llm-gateway): re-enable claude-fable-5#67267

Merged
charlesvien merged 2 commits into
masterfrom
feat/re-enable-claude-fable-5
Jul 1, 2026
Merged

feat(llm-gateway): re-enable claude-fable-5#67267
charlesvien merged 2 commits into
masterfrom
feat/re-enable-claude-fable-5

Conversation

@charlesvien

@charlesvien charlesvien commented Jul 1, 2026

Copy link
Copy Markdown
Member

Problem

we're so back https://x.com/AnthropicAI/status/2072106151890809341

Changes

  • Re-add claude-fable-5 to the shared _POSTHOG_CODE_AGENT_MODELS allowlist (covers posthog_code and slack_app) and to background_agents, so it lists in /{product}/v1/models and requests for it are accepted again
  • Restore its Bedrock mapping so the Bedrock provider path resolves us./eu.anthropic.claude-fable-5
  • No cost change needed: the LiteLLM cost-override bridge entry was intentionally kept in fix(llm-gateway): disable claude-fable-5 #63382, so billing and resolution never went away

How did you test this code?

Agent-authored (Claude Code). I did not do manual testing. Ran the three affected llm-gateway suites locally with the service venv:

pytest tests/test_bedrock.py tests/test_model_cost_overrides.py tests/test_product_config.py -> 223 passed.

The restored cases re-assert fable-5's allowlist membership (posthog_code, background_agents), its Bedrock inference-profile mapping, and its presence in the /v1/models listing. I deleted the disable-era test_posthog_code_rejects_claude_fable_5_via_bedrock_provider guard: restoring the Bedrock mapping is exactly what makes fable resolvable through that path again, so the guard no longer holds.

@charlesvien charlesvien self-assigned this Jul 1, 2026
@charlesvien charlesvien marked this pull request as ready for review July 1, 2026 07:43
@charlesvien charlesvien added the stamphog Request AI approval (no full review) label Jul 1, 2026
@charlesvien charlesvien requested a review from a team July 1, 2026 07:44
@assign-reviewers-posthog assign-reviewers-posthog Bot requested a review from a team July 1, 2026 07:44

@richardsolomou richardsolomou left a comment

Copy link
Copy Markdown
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

never stamped anything faster

@stamphog stamphog Bot left a comment

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This PR makes a behavioral change to the LLM gateway's model access control — re-enabling claude-fable-5 for the posthog_code product — but the author is not on the owning team (@PostHog/team-ai-gateway) and has zero reviews from that team. Non-owner behavioral changes to business logic require a human review from the owning team.

@stamphog stamphog Bot removed the stamphog Request AI approval (no full review) label Jul 1, 2026
@greptile-apps

greptile-apps Bot commented Jul 1, 2026

Copy link
Copy Markdown
Contributor

Reviews (1): Last reviewed commit: "re-enable claude-fable-5 in llm gateway" | Re-trigger Greptile

Comment thread services/llm-gateway/tests/test_model_cost_overrides.py
@charlesvien charlesvien merged commit cb6cfcd into master Jul 1, 2026
178 checks passed
@charlesvien charlesvien deleted the feat/re-enable-claude-fable-5 branch July 1, 2026 19:39
@deployment-status-posthog

deployment-status-posthog Bot commented Jul 1, 2026

Copy link
Copy Markdown

Deploy status

Environment Status Deployed At Workflow
dev ✅ Deployed 2026-07-01 20:04 UTC Run
prod-us ✅ Deployed 2026-07-01 20:20 UTC Run
prod-eu ✅ Deployed 2026-07-01 20:25 UTC Run

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants