v1.0.0 release broke model endpoints for API v1beta

[matteopariset]

v1.0.0 release broke model endpoints for API v1beta

---

Describe the bug
After the release of v1.0.0, requests using :contentReference[oaicite:0]{index=0} via the v1beta API version started failing when using certain models (e.g., gemini-3-flash). This issue did not occur prior to the release.

---

To Reproduce
Steps to reproduce the behavior:

1. Use a TypeScript backend with an LlmAgent configured as follows:

   export function createRootAgent(storeReference: string) {
     return new LlmAgent({
       model: MODEL_ID,
       name: "RagAgent",
       description:
         "Agent to perform retrieval and reasoning from the file search store",
       instruction: RAG_AGENT_INSTRUCTION,
       outputSchema: RAG_OUTPUT_SCHEMA,
       tools: [new FileSearchTool(storeReference)],
     });
   }

2. Set MODEL_ID = "models/gemini-3-flash"

3. Execute a request using the v1beta API

4. Observe the error:

   {
     "errorMessage": "models/gemini-3-flash is not found for API version v1beta, or is not supported for generateContent. Call ListModels to see the list of available models and their supported methods."
   }

---

Expected behavior
models/gemini-3-flash should continue to work with the v1beta API as it did prior to the v1.0.0 release, or there should be clear migration guidance if support has changed.

---

Screenshots
N/A (error message included above)

---

Desktop (please complete the following information):

• OS: macOS (assumed, can adjust if needed)
• TS version/environment: Node.js (TypeScript backend)
• ADK version (see maven dependency): v1.0.0

---

Additional context

• The issue started occurring approximately 3 hours after the v1.0.0 release.

• No changes were made to the application code or configuration during that time.

• This suggests a breaking change in model availability or API compatibility for v1beta.

• It is unclear whether:

  • gemini-3-flash was removed from v1beta, or
  • model naming / endpoint requirements changed in v1.0.0.

---

Questions

• Has gemini-3-flash been deprecated or moved to a different API version?
• Is there a migration path for v1beta users?
• Should we switch to v1 or a different model identifier?

---

[Varun-S10]

Hi @matteopariset, We have not been able to reproduce this issue on our end. To help us investigate further, could you please share the code where you encountered this problem? Thank you.