chore: update model catalog from bot issues

[github-actions]

Automated daily batch of model catalog updates from bot issues.

Included issues

• Closes [BOT ISSUE] Together: update stale Qwen/Qwen3.7-Max pricing ($1.25/$3.75 → $2.00/$6.00) #1245: [BOT ISSUE] Together: update stale Qwen/Qwen3.7-Max pricing ($1.25/$3.75 → $2.00/$6.00)
• Closes [BOT ISSUE] Together: add missing together provider to deepseek-ai/DeepSeek-V4.1-Flash #1246: [BOT ISSUE] Together: add missing together provider to deepseek-ai/DeepSeek-V4.1-Flash

Summary

Issue	Provider	Primary model	Changed models	Added models	Updated models	Verification sources
#1245	together	Qwen/Qwen3.7-Max	Qwen/Qwen3.7-Max	None	Qwen/Qwen3.7-Max	1 2
#1246	together	deepseek-ai/DeepSeek-V4.1-Flash	deepseek-ai/DeepSeek-V4.1-Flash	None	deepseek-ai/DeepSeek-V4.1-Flash	1 2

Verified metadata

#1245: [BOT ISSUE] Together: update stale Qwen/Qwen3.7-Max pricing ($1.25/$3.75 → $2.00/$6.00)

Model	Display name	Parent	Providers	Format	Flavor	Token limits	Pricing	Lifecycle
Qwen/Qwen3.7-Max	Qwen3.7 Max		together	openai	chat	input=1000000, output=not provided	in/out=2/6 per 1M; cache read=0.25 per 1M	reasoning=true

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.

Model	Field	Proposed update	sync_models	sync_models source models
Qwen/Qwen3.7-Max	input_cost_per_mil_tokens	2	2.5	together_ai/Qwen/Qwen3.7-Max
Qwen/Qwen3.7-Max	output_cost_per_mil_tokens	6	7.5	together_ai/Qwen/Qwen3.7-Max
Qwen/Qwen3.7-Max	input_cache_read_cost_per_mil_tokens	0.25	0.5	together_ai/Qwen/Qwen3.7-Max

#1246: [BOT ISSUE] Together: add missing together provider to deepseek-ai/DeepSeek-V4.1-Flash

Model	Display name	Parent	Providers	Format	Flavor	Token limits	Pricing	Lifecycle
deepseek-ai/DeepSeek-V4.1-Flash	DeepSeek V4.1 Flash		baseten, fireworks, together	openai	chat	input=1048576, output=not provided	in/out=0.3/1.2 per 1M; cache read=0.03 per 1M	multimodal=true; reasoning=true

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.

Model	Field	Proposed update	sync_models	sync_models source models
deepseek-ai/DeepSeek-V4.1-Flash	catalog entry	present	missing	None

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Actions	Updated
ai-proxy	Ready	Preview	Sep 14, 2026 10:19am UTC

[github-actions]

Codex (@codex) review

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 0e4543b625

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".

[chatgpt-codex-connector]

Keep provider routing out of the deprecated proxy

Remove this AvailableEndpointTypes update: it changes getDirectModelEndpointTypes so the deprecated proxy advertises Together routing for this model, while repository guidance places routing changes in ../gateway/ and makes an exception only for model_list.json. Keeping the catalog update is valid, but updating this map perpetuates routing behavior in the wrong implementation while leaving the active gateway unchanged.

AGENTS.md reference: AGENTS.md:L6-L15

Useful? React with 👍 / 👎.