feat(deepinfra/deepseek-ai/DeepSeek-V4-Pro-0813): add new models [bot] - #2170
feat(deepinfra/deepseek-ai/DeepSeek-V4-Pro-0813): add new models [bot]#2170models-bot[bot] wants to merge 1 commit into
Conversation
![cursor[bot]](https://avatars.githubusercontent.com/in/1210556?s=60&v=4){kind=small}
There was a problem hiding this comment.
Cursor Bugbot has reviewed your changes using default effort and found 3 potential issues.
❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.
Reviewed by Cursor Bugbot for commit ca8a09a. Configure here.
| limits: | ||
| context_window: 1048576 | ||
| max_tokens: 1048576 | ||
| mode: unknown |
There was a problem hiding this comment.
Wrong primary mode set
High Severity
mode is set to unknown for deepseek-ai/DeepSeek-V4-Pro-0813, a chat LLM. Sibling DeepInfra entries such as DeepSeek-V4-Pro.yaml and DeepSeek-V4-Flash-0731.yaml use mode: chat. Downstream consumers that key off mode will not treat this model as a chat model.
Reviewed by Cursor Bugbot for commit ca8a09a. Configure here.
| output_cost_per_token: 2.6e-6 | ||
| region: "*" | ||
| features: | ||
| - prompt_caching |
There was a problem hiding this comment.
Missing tool and JSON features
Medium Severity
features only lists prompt_caching. The same DeepInfra DeepSeek-V4-Pro family documents and sibling configs also advertise function_calling and json_output. Clients that gate tools or structured output on these flags will disable supported capabilities for this model.
Reviewed by Cursor Bugbot for commit ca8a09a. Configure here.
| context_window: 1048576 | ||
| max_tokens: 1048576 | ||
| mode: unknown | ||
| model: deepseek-ai/DeepSeek-V4-Pro-0813 |
There was a problem hiding this comment.
Missing thinking capability flag
Medium Severity
The new config never sets thinking: true. DeepSeek-V4-Pro.yaml sets it, and DeepInfra documents reasoning/thinking modes for this family. Consumers that rely on thinking will not recognize this model as a reasoning model.
Reviewed by Cursor Bugbot for commit ca8a09a. Configure here.
1 participant
Auto-generated by model-addition-agent for
deepinfra/deepseek-ai/DeepSeek-V4-Pro-0813.Note
Low Risk
Adds static provider metadata only; no application or inference logic changes.
Overview
Adds a new DeepInfra provider catalog file for
deepseek-ai/DeepSeek-V4-Pro-0813, enabling routing and cost accounting for that endpoint.The entry defines 1M-token context and max output limits, prompt caching (with separate cache-read input pricing), and per-token input/output costs.
modeis set tounknown, and the stub omits fields that fuller sibling configs often include (e.g.chatmode, modalities, sources, status).Reviewed by Cursor Bugbot for commit ca8a09a. Bugbot is set up for automated code reviews on this repo. Configure here.