mirror of
https://github.com/BerriAI/litellm.git
synced 2026-10-09 03:18:44 +00:00
3d1f7f3544
The weekly auto_update_price_and_context_window job crashes with KeyError: 'output' when the Vercel AI Gateway /v1/models response includes models whose pricing lacks the 'output' key. The current upstream payload has ~88 such rows: embedding models (input-only), image generators, video generators, and a handful with empty pricing. Skip rows that have neither input nor output token pricing (image and video models do not map to token-based cost fields anyway) and fall back to 0.0 for the standard chat/embedding rows so embeddings get included with input-only pricing instead of aborting the whole sync. Add a regression test covering the embedding-only, image-only, video-only, empty-pricing, and standard chat-with-cache shapes. Co-authored-by: Krrish Dholakia <krrish-berri-2@users.noreply.github.com> |
||
|---|---|---|
| .. | ||
| codeql | ||
| ISSUE_TEMPLATE | ||
| observatory | ||
| screenshots | ||
| scripts | ||
| workflows | ||
| dependabot.yaml | ||
| deploy-to-aws.png | ||
| FUNDING.yml | ||
| pull_request_template.md | ||
| template.yaml | ||
