chore: update model catalog from bot issues#913
Conversation
github-actions
Bot
requested review from
Alex Z (CLowbrow),
aswink,
Caitlin Pinn (cpinn),
Erin McNulty (erin2722) and
Ken Jiang (knjiang)
|
Codex (@codex) review |
|
The latest updates on your projects. Learn more about Vercel for GitHub.
|
![chatgpt-codex-connector[bot]](https://avatars.githubusercontent.com/in/1144995?s=60&v=4){kind=small}
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 63e7c8f6f0
ℹ️ About Codex in GitHub
Codex has been enabled to automatically review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "Codex (@codex) review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
When you sign up for Codex through ChatGPT, Codex can also answer questions or update the PR, like "Codex (@codex) address that feedback".
| "bedrock" | ||
| ] | ||
| }, | ||
| "global.anthropic.claude-opus-4-6-v1:0": { |
There was a problem hiding this comment.
For Claude Opus 4.6, the AWS Bedrock model card lists the runtime and global inference IDs without the :0 suffix (anthropic.claude-opus-4-6-v1, us.anthropic.claude-opus-4-6-v1, global.anthropic.claude-opus-4-6-v1) (AWS docs). fetchBedrock passes the catalog model name directly as modelId, so this newly added global...:0 entry advertises an identifier that Bedrock rejects; requests choosing this global model will fail until the no-suffix ID is cataloged instead.
Useful? React with 👍 / 👎.
| "reasoning": true, | ||
| "reasoning_budget": true, | ||
| "parent": "anthropic.claude-opus-4-6-v1:0", | ||
| "max_input_tokens": 200000, |
There was a problem hiding this comment.
The AWS model card lists a 1M-token context window for Claude Opus 4.6 (AWS docs), but this new global Bedrock entry caps max_input_tokens at 200,000. Any UI or validation that relies on this catalog will under-report or restrict the model and prevent users from using the documented long-context capacity.
Useful? React with 👍 / 👎.
Verified by real Bedrock invocation with the CI credentials (us-east-1): - anthropic.claude-opus-4-6-v1:0 / us. / global. (the PR's ids) all return HTTP 400 "The provided model identifier is invalid." (Codex P1). - The correct ids drop the :0 suffix: us.anthropic.claude-opus-4-6-v1 and global.anthropic.claude-opus-4-6-v1 return HTTP 200 and stream real content both non-streaming and streaming. (The base anthropic.claude-opus-4-6-v1 is not on-demand invocable, same as anthropic.claude-sonnet-4-6 — callers use the us./global. inference-profile ids.) Also raise max_input_tokens 200000 -> 1000000 to match Claude Opus 4.6's 1M context window (the anthropic-direct claude-opus-4-6 entry; Codex P2). Renamed the keys, parent references, and index.ts mappings. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
…ues-2026-07-04 # Conflicts: # packages/proxy/schema/index.ts # packages/proxy/schema/model_list.json
1 participant
Automated daily batch of model catalog updates from bot issues.
Included issues
Summary
anthropic.claude-opus-4-6-v1:0us.anthropic.claude-opus-4-6-v1:0anthropic.claude-opus-4-6-v1:0us.anthropic.claude-opus-4-6-v1:02
global.anthropic.claude-opus-4-6-v1:0global.anthropic.claude-opus-4-6-v1:0Verified metadata
#911: [BOT ISSUE] Bedrock: add missing available_providers on anthropic.claude-opus-4-6-v1:0 and us variant
Verification notes
No LLM verification step ran; model metadata was already complete in the issue.
sync_models vs proposed update
sync_models cross-check found differences. Official provider verification was used for the applied values, and sync_models discrepancies are listed below for review.
#912: [BOT ISSUE] Bedrock: add missing global.anthropic.claude-opus-4-6-v1:0 cross-region entry
Verification notes
No LLM verification step ran; model metadata was already complete in the issue.
sync_models vs proposed update
sync_models cross-check found differences. Official provider verification was used for the applied values, and sync_models discrepancies are listed below for review.