Skip to content
This repository was archived by the owner on Jul 17, 2026. It is now read-only.

feat: add jimeng-4.7 image model and Seedance 2.0 variants (mini/fast-vip/vip) - #171

Open
CNQQC wants to merge 1 commit into
iptag:mainfrom
CNQQC:feat/add-jimeng-4.7-and-seedance-2.0-variants
Open

feat: add jimeng-4.7 image model and Seedance 2.0 variants (mini/fast-vip/vip)#171
CNQQC wants to merge 1 commit into
iptag:mainfrom
CNQQC:feat/add-jimeng-4.7-and-seedance-2.0-variants

Conversation

@CNQQC

@CNQQC CNQQC commented Jul 2, 2026

Copy link
Copy Markdown

Summary

The official Jimeng site has shipped several models that this wrapper did not yet map. This PR adds them.

Image

  • jimeng-4.7high_aes_general_v43 (added to IMAGE_MODEL_MAP and IMAGE_MODEL_MAP_ASIA)
  • jimeng-4.7 added to the intelligent_ratio allow-list

Video (China site, Seedance 2.0 family)

model req_key benefit_type
jimeng-video-seedance-2.0-mini dreamina_seedance_40_mini seedance_20_mini_720p_output
jimeng-video-seedance-2.0-fast-vip dreamina_seedance_40_vision seedance_20_fast_720p_output
jimeng-video-seedance-2.0-vip dreamina_seedance_40_pro_vision seedance_20_pro_720p_output

Note: in getVideoBenefitType() the 40_pro_vision branch must be checked before 40_pro and 40_vision, because the VIP req_key (dreamina_seedance_40_pro_vision) contains all three substrings.

The pre-existing jimeng-video-seedance-2.0 (dreamina_seedance_40_pro) maps to the site's base "Seedance 2.0" and is unchanged.

Docs

  • Updated the supported-model lists and the intelligent_ratio note in both README.md and README.CN.md.

Testing

  • jimeng-4.7: verified end-to-end — valid 2048×2048 PNG.
  • seedance-2.0-mini and seedance-2.0-fast-vip: verified end-to-end via /v1/videos/generations — valid MP4 output.
  • seedance-2.0-vip: request accepted by the backend and generation started (req_key confirmed valid).

Model ids were obtained from the aigc_draft/generate request model_id / benefit_type fields.

…-vip/vip)

Image:
- Add jimeng-4.7 (high_aes_general_v43) to IMAGE_MODEL_MAP and IMAGE_MODEL_MAP_ASIA
- Allow intelligent_ratio for jimeng-4.7

Video (CN, Seedance 2.0 family):
- Add jimeng-video-seedance-2.0-mini    -> dreamina_seedance_40_mini
- Add jimeng-video-seedance-2.0-fast-vip -> dreamina_seedance_40_vision
- Add jimeng-video-seedance-2.0-vip      -> dreamina_seedance_40_pro_vision
- Map their benefit_type in getVideoBenefitType(); the 40_pro_vision branch is
  ordered before 40_pro/40_vision because the VIP req_key contains all three
  substrings

Docs: update model lists and intelligent_ratio notes in README.md / README.CN.md

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant