luna/oh-my-opencode
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

fix(provider): use gateway-specific model IDs for vercel transform

Browse Source 
The gateway uses google/gemini-3-flash (no -preview suffix) unlike the
direct Google API. Give the vercel provider its own transform instead
of delegating to sub-provider transforms blindly.
This commit is contained in:
Matan Kushner
2026-04-13 12:51:56 +09:00
parent 6f72ec754d
commit 542dc890f9
3 changed files with 21 additions and 12 deletions
Download Patch File Download Diff File Expand all files Collapse all files
src/cli/provider-model-id-transform.test.ts
+3 -3
View File
@@ -247,13 +247,13 @@ describe("transformModelForProvider", () => {
expect(result).toBe("google/gemini-3.1-pro-preview")
})
test("prepends google/ and applies google transform for gemini-3-flash", () => {
test("prepends google/ without -preview for gemini-3-flash", () => {
// #given vercel provider and gemini-3-flash model
// #when transformModelForProvider is called
const result = transformModelForProvider("vercel", "gemini-3-flash")
// #then should produce google/gemini-3-flash-preview
expect(result).toBe("google/gemini-3-flash-preview")
// #then should produce google/gemini-3-flash (gateway does not use -preview for this model)
expect(result).toBe("google/gemini-3-flash")
})
test("prepends xai/ for grok models", () => {