Add gemini-3.8-flash to Gemini CUA, HyperAgent, and Browser Use LLM types - #113
Conversation
🤖 Devin AI EngineerI'll be helping with this pull request! Here's what you should know: ✅ I will automatically:
Note: I can only respond to comments from users who have write access to this repository. ⚙️ Control Options:
|
|
| Filename | Overview |
|---|---|
| src/types/constants.ts | Adds the new model identifier to all three intended TypeScript unions without changing runtime behavior. |
| src/tools/schema.ts | Adds the model identifier to the shared Browser Use schema enum, resolving the previously reported strict-validation mismatch. |
Reviews (2): Last reviewed commit: "Add gemini-3.8-flash to browser-use tool..." | Re-trigger Greptile
| | "gemini-2.0-flash" | ||
| | "gemini-2.5-flash"; | ||
| | "gemini-2.5-flash" | ||
| | "gemini-3.8-flash"; |
There was a problem hiding this comment.
Browser-use schema rejects new model
When an OpenAI or Anthropic strict tool definition selects gemini-3.8-flash for llm, plannerLlm, or pageExtractionLlm, BrowserUseLlm accepts the value but BROWSER_USE_LLM_SCHEMA omits it, causing strict validation to reject or exclude the newly supported model.
Prompt To Fix With AI
This is a comment left during a code review.
Path: src/types/constants.ts
Line: 52
Comment:
**Browser-use schema rejects new model**
When an OpenAI or Anthropic strict tool definition selects `gemini-3.8-flash` for `llm`, `plannerLlm`, or `pageExtractionLlm`, `BrowserUseLlm` accepts the value but `BROWSER_USE_LLM_SCHEMA` omits it, causing strict validation to reject or exclude the newly supported model.
---
For each issue above, determine whether it is valid and should be fixed. If so, fix it directly.
Summary
Adds
"gemini-3.8-flash"toGeminiComputerUseLlm,HyperAgentLlm, andBrowserUseLlm. Server support: hyperbrowserai/browserctrl#1385 (it is now the server-side default for Gemini Computer Use).Changes
src/types/constants.ts: new"gemini-3.8-flash"literal in the three LLM union types.Validation
corepack yarn buildLink to Devin session: https://app.devin.ai/sessions/7dccc05e05b04c63bd5760859b161275
Open in Devin Desktop: https://app.devin.ai/desktop/session/7dccc05e05b04c63bd5760859b161275?variant=devin
Requested by: @shrisukhani
Note
Low Risk
Additive type and JSON-schema enum updates only; no auth, data, or execution logic changes.
Overview
Adds
gemini-3.8-flashas a selectable model across client types and tool schemas so agents can target the newer Gemini flash model (aligned with server default in browserctrl#1385).In
src/types/constants.ts, the literal is added toBrowserUseLlm,HyperAgentLlm, andGeminiComputerUseLlm. Insrc/tools/schema.ts, the same value is included inBROWSER_USE_LLM_SCHEMAforllm,plannerLlm, andpageExtractionLlmon browser-use tools. Existing defaults remaingemini-2.0-flashfor browser use.Reviewed by Cursor Bugbot for commit f223684. Bugbot is set up for automated code reviews on this repo. Configure here.