feat(together-ai/black-forest-labs/FLUX-3): add new models [bot] - #2043
Conversation
|
/test-models |
Gateway test results
Skipped (1)
Skip reason |
|
/test-models |
Gateway test results
Skipped (1)
Skip reason |
|
/test-models |
Gateway test results
Skipped (1)
Skip reason |
There was a problem hiding this comment.
Cursor Bugbot has reviewed your changes using default effort and found 1 potential issue.
❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.
Reviewed by Cursor Bugbot for commit 1be7601. Configure here.
| - image | ||
| mode: video | ||
| model: black-forest-labs/FLUX-3 | ||
| provisioning: serverless |
There was a problem hiding this comment.
Missing reasoning_effort param removal
Medium Severity
This video model omits removeParams for reasoning_effort, so it inherits the Together AI provider default chat parameter. Peer video and newer FLUX configs strip that param; without the override, callers may see or send an unsupported reasoning setting for video generation.
Reviewed by Cursor Bugbot for commit 1be7601. Configure here.
|
This is a video generation model with per-video pricing ($0.17) expressed via a YAML comment. The per-token cost schema can't represent this natively — is there a plan for how the gateway should handle video-gen pricing? |


Auto-generated by model-addition-agent for
together-ai/black-forest-labs/FLUX-3.Note
Low Risk
Metadata-only addition with no runtime or billing logic changes; placeholder costs may mislead cost estimates until schema supports per-video pricing.
Overview
Adds a new Together AI provider catalog entry for
black-forest-labs/FLUX-3, registering it as a preview, serverless video model with text/image/video inputs and video/audio/image outputs.Cost fields use zero token placeholders with inline notes that Together’s published $0.17 per video pricing cannot be represented in the current schema. Sources link to Together and BFL documentation.
Reviewed by Cursor Bugbot for commit 1be7601. Bugbot is set up for automated code reviews on this repo. Configure here.