backend new models check - byteplus - 2026-09-16
- Dominant language
- TypeScript
- Stars
- 43.5k
- Forks
- 4.1k
- Avg merge
- 1d 3h
- Merged PRs (30d)
- 226
Description
Files checked: `src/backend/drivers/ai-chat/providers/byteplus/models.ts`, `src/backend/drivers/ai-image/providers/byteplus/models.ts`, `src/backend/drivers/ai-video/providers/byteplus/models.ts`
**chat (3)**
- `deepseek-v4-flash-ga-260731` | family: `deepseek-v4-flash-ga` | notes: new model — separate "-ga" (general availability) family from backend's `deepseek-v4-flash-260425`
- `deepseek-v4-pro-ga-260813` | family: `deepseek-v4-pro-ga` | notes: new model — separate "-ga" family from backend's `deepseek-v4-pro-260425`
- `glm-5-3-flash-260828` | family: `glm-5-3-flash` | notes: new model, no existing glm-5-3 family in backend
**image (0)** — all upstream seedream* families (`dola-seedream-5-0-pro-260628`, `seedream-5-0-260128`, `seedream-4-5-251128`, `seedream-4-0-250828`) match backend ids/aliases and dates exactly.
**video (0)** — all upstream seedance*/dreamina-seedance* families match backend ids and dates exactly.
skipped (unsupported modality): 4 — `skylark-embedding-vision-251215`, `skylark-embedding-vision-250615` (embedding), `Hyper3d-Rodin-Gen2` (hyper3d), `Hitem3d-2.0` (hitem3d)
Note: byteplus family "seed-2-0-code" (upstream) vs backend's "seed-2-0-code-preview" naming is a cosmetic family-label difference only — same id/date (`seed-2-0-code-preview-260328`) present in both, not counted as new. All other byteplus chat families have their newest upstream date already matching the backend's newest date — no version-bump gaps found there.
Contributor guide
Research direction
Start with src/backend/drivers/ai-chat/providers/byteplus/models.ts and compare its existing BytePlus chat model entries with the three models listed in the issue. Check the image and video provider files for context, then verify that the three new chat models are represented without changing already-matching entries; done means the listed gaps are covered and unsupported modalities remain skipped.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- typescript
- Domain
- ai, backend
- Issue type
- Feature
- Difficulty
- 2/5
- Estimated time
- 1-3 hours
- Activity status
- Active
- Clarity
- Clearly specified
- Newbie friendliness
- 85/100