chore: update model catalog from bot issues#876
Conversation
github-actions
Bot
requested review from
Alex Z (CLowbrow),
aswink,
Caitlin Pinn (cpinn),
Erin McNulty (erin2722) and
Ken Jiang (knjiang)
|
The latest updates on your projects. Learn more about Vercel for GitHub.
|
|
Codex (@codex) review |
![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: d9503dbdc0
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".
| "locations": [ | ||
| "global" | ||
| ], |
There was a problem hiding this comment.
When a Vertex secret is configured with a regional or multi-region location, this locations override makes resolveVertexLocation choose modelSpec.locations before the secret metadata (packages/proxy/src/proxy.ts:2645), so every Vertex call to Sonnet 4.6 is forced to /locations/global/. Google documents Sonnet 4.6 as available in us-east5, europe-west1, asia-southeast1, and global (https://docs.cloud.google.com/gemini-enterprise-agent-platform/models/partner-models/claude/sonnet-4-6#supported-regions), so customers relying on regional routing/data residency lose their configured location unless this is removed or all supported locations are represented without overriding metadata.
Useful? React with 👍 / 👎.
| "available_providers": [ | ||
| "google" | ||
| ] |
There was a problem hiding this comment.
For Vertex-only installations, this new model will not be selectable because it advertises only the google provider and has no publisher fallback, while Google Cloud lists gemini-3.1-flash-image as available on the global endpoint (https://docs.cloud.google.com/gemini-enterprise-agent-platform/models/gemini/3-1-flash-image#supported-regions). The adjacent stable image model (gemini-3-pro-image) includes both google and vertex plus a publishers/google/models/... fallback, so this omission prevents routing through supported Vertex credentials.
Useful? React with 👍 / 👎.
github-actions
Bot
force-pushed
the
chore/autofix-bot-issues-2026-06-26
branch
from
d9503db to
55006dc
Compare
Codex Review ResponseIssue 1: Preserve Vertex region selection for Claude 4.6 (P2)Fixed. Removed the Issue 2: Expose Gemini 3.1 Flash Image on Vertex (P2)Fixed. The
|
The bot bumped publishers/anthropic/models/claude-sonnet-4-6 max_output_tokens to 128000, but Claude Sonnet 4.6 has a 64000 output cap (LiteLLM lists 64000 for both claude-sonnet-4-6 and vertex_ai/claude-sonnet-4-6, and the non-vertex catalog entry is 64000). The max_input bump to 1000000 is correct (Sonnet 4.6 is 1M context); only the output value was wrong. Keep the two provider variants consistent at 1M / 64000. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
1 participant
Automated daily batch of model catalog updates from bot issues.
Included issues
Summary
claude-sonnet-4-6anthropic.claude-sonnet-4-6us.anthropic.claude-sonnet-4-6global.anthropic.claude-sonnet-4-6publishers/anthropic/models/claude-sonnet-4-6claude-sonnet-4-6anthropic.claude-sonnet-4-6us.anthropic.claude-sonnet-4-6global.anthropic.claude-sonnet-4-6publishers/anthropic/models/claude-sonnet-4-62
3
google/gemma-4-31B-itQwen/Qwen3.5-9Bmeta-llama/Llama-3.3-70B-Instruct-Turbogoogle/gemma-4-31B-itQwen/Qwen3.5-9Bmeta-llama/Llama-3.3-70B-Instruct-Turbo2
gpt-5gpt-5.4-minigpt-5.4-nanogpt-5gpt-5.4-minigpt-5.4-nano2
3
4
pixtral-large-latestpixtral-12b-2409mistral-saba-latestpixtral-large-latestpixtral-12b-2409mistral-saba-latest2
meta-llama/llama-4-maverick-17b-128e-instructmoonshotai/kimi-k2-instruct-0905meta-llama/llama-4-maverick-17b-128e-instructmoonshotai/kimi-k2-instruct-09052
gpt-5-progpt-5-pro-2025-10-06gpt-5-progpt-5-pro-2025-10-062
chat-latestchat-latest2
codestral-2508voxtral-small-2507codestral-2508voxtral-small-25072
grok-4.20-multi-agent-0309grok-4.20-multi-agent-03092
gemini-3.1-flash-imagegemini-3.1-flash-image2
llama-3.1-8b-instantllama-3.3-70b-versatileqwen/qwen3-32bllama-3.1-8b-instantllama-3.3-70b-versatileqwen/qwen3-32b2
meta-llama/llama-4-scout-17b-16e-instructgemma2-9b-itllama3-70b-8192meta-llama/llama-4-scout-17b-16e-instructgemma2-9b-itllama3-70b-81922
llama-4-scout-17b-16e-instructllama3.3-70bllama3.1-8bllama-4-scout-17b-16e-instructllama3.3-70bllama3.1-8b2
llama3-8b-8192mistral-saba-24bdeepseek-r1-distill-llama-70bllama3-8b-8192mistral-saba-24bdeepseek-r1-distill-llama-70b2
3
deepseek-r1-distill-llama-70b-specdecdeepseek-r1-distill-qwen-32bqwen-2.5-32bdeepseek-r1-distill-llama-70b-specdecdeepseek-r1-distill-qwen-32bqwen-2.5-32b2
qwen-2.5-coder-32bqwen-qwq-32bllama-guard-3-8bqwen-2.5-coder-32bqwen-qwq-32bllama-guard-3-8b2
llama-3.2-90b-vision-previewllama-3.2-11b-vision-previewllama-3.2-3b-previewllama-3.2-90b-vision-previewllama-3.2-11b-vision-previewllama-3.2-3b-preview2
llama-3.2-1b-previewllama-3.2-1b-preview2
Verified metadata
#851: [BOT ISSUE] Anthropic: claude-sonnet-4-6 max_output_tokens should be 128000, not 64000
Verification notes
Verification
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.
#852: [BOT ISSUE] Together: stale pricing for google/gemma-4-31B-it, Qwen/Qwen3.5-9B, and meta-llama/Llama-3.3-70B-Instruct-Turbo
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.
#853: [BOT ISSUE] OpenAI: stale max_input_tokens for gpt-5, gpt-5.4-mini, and gpt-5.4-nano (272000 → 400000)
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.
#854: [BOT ISSUE] Mistral: missing deprecation flags for pixtral-large-latest, pixtral-12b-2409, and mistral-saba-latest
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.
#855: [BOT ISSUE] Groq: missing deprecation flags for meta-llama/llama-4-maverick-17b-128e-instruct and moonshotai/kimi-k2-instruct-0905
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.
#859: [BOT ISSUE] OpenAI: gpt-5-pro and gpt-5-pro-2025-10-06 stale max_input_tokens (128000 → 400000)
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.
#860: [BOT ISSUE] OpenAI: chat-latest stale max_input_tokens (128000 → 400000) and max_output_tokens (16384 → 128000)
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.
#862: [BOT ISSUE] Mistral: codestral-2508 stale max_input_tokens (256000 → 128000) and voxtral-small-2507 stale output pricing ($0.4 → $0.3)
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.
#866: [BOT ISSUE] xAI: missing canonical model ID grok-4.20-multi-agent-0309
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.
#867: [BOT ISSUE] Google: missing stable model gemini-3.1-flash-image
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.
#868: [BOT ISSUE] Groq: missing deprecation_date for llama-3.1-8b-instant, llama-3.3-70b-versatile, and qwen/qwen3-32b
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.
#869: [BOT ISSUE] Groq: missing deprecation flags for meta-llama/llama-4-scout-17b-16e-instruct, gemma2-9b-it, and llama3-70b-8192
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.
#870: [BOT ISSUE] Cerebras: missing deprecation flags for llama-4-scout-17b-16e-instruct, llama3.3-70b, and llama3.1-8b
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.
#871: [BOT ISSUE] Groq: missing deprecation flags for llama3-8b-8192, mistral-saba-24b, and deepseek-r1-distill-llama-70b
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.
#872: [BOT ISSUE] Groq: missing deprecation flags for deepseek-r1-distill-llama-70b-specdec, deepseek-r1-distill-qwen-32b, and qwen-2.5-32b
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.
#873: [BOT ISSUE] Groq: missing deprecation flags for qwen-2.5-coder-32b, qwen-qwq-32b, and llama-guard-3-8b
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.
#874: [BOT ISSUE] Groq: missing deprecation flags for llama-3.2-90b-vision-preview, llama-3.2-11b-vision-preview, and llama-3.2-3b-preview
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.
#875: [BOT ISSUE] Groq: missing deprecation flag for llama-3.2-1b-preview
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.