received 1008 (policy violation) <Live Model Name> is not found for API version v1alpha, or is not supported for bidiGenerateContent

[ravi-incred]

Describe the bug

I had built an ADK Agent using Google Gemini LIVE Models for audio-to-audio conversation and suddenly my ADK Agent stoped working and later reading the server logs I found out this message:

Connection closed: received 1008 (policy violation) models/gemini-live-2.5-flash-preview is not found for API version v1alpha, or is not supported for bidiGenerateContent; then sent 1008 (policy violation) models/gemini-live-2.5-flash-native-audio is not found for API version v1alpha, or is not supported for bidiGenerateContent.
WebSocket error: received 1008 (policy violation) models/gemini-live-2.5-flash-preview is not found for API version v1alpha, or is not supported for bidiGenerateContent; then sent 1008 (policy violation) models/gemini-live-2.5-flash-native-audio is not found for API version v1alpha, or is not supported for bidiGenerateContent

Note - I am using Google ADK with Google AI Studio setup
my .env look like this

GOOGLE_GENAI_USE_VERTEXAI=FALSE
GOOGLE_API_KEY=<KEY>

Additional Context

I tried 3 models hoping that maybe its a model issue

• models/gemini-live-2.5-flash-native-audio
• models/gemini-live-2.5-flash-preview
• models/gemini-live-2.5-flash
  But got the same response

Connection closed: received 1008 (policy violation) models/gemini-live-2.5-flash-native-audio is not found for API version v1alpha, or is not supported for bidiGenerateContent; then sent 1008 (policy violation) models/gemini-live-2.5-flash-native-audio is not found for API version v1alpha, or is not supported for bidiGenerateContent.
WebSocket error: received 1008 (policy violation) models/gemini-live-2.5-flash-native-audio is not found for API version v1alpha, or is not supported for bidiGenerateContent; then sent 1008 (policy violation) models/gemini-live-2.5-flash-native-audio is not found for API version v1alpha, or is not supported for bidiGenerateContent


Connection closed: received 1008 (policy violation) models/gemini-live-2.5-flash-preview is not found for API version v1alpha, or is not supported for bidiGenerateContent; then sent 1008 (policy violation) models/gemini-live-2.5-flash-native-audio is not found for API version v1alpha, or is not supported for bidiGenerateContent.
WebSocket error: received 1008 (policy violation) models/gemini-live-2.5-flash-preview is not found for API version v1alpha, or is not supported for bidiGenerateContent; then sent 1008 (policy violation) models/gemini-live-2.5-flash-native-audio is not found for API version v1alpha, or is not supported for bidiGenerateContent

onnection closed: received 1008 (policy violation) models/gemini-live-2.5-flash is not found for API version v1alpha, or is not supported for bidiGenerateContent; then sent 1008 (policy violation) models/gemini-live-2.5-flash-native-audio is not found for API version v1alpha, or is not supported for bidiGenerateContent.
WebSocket error: received 1008 (policy violation) models/gemini-live-2.5-flash is not found for API version v1alpha, or is not supported for bidiGenerateContent; then sent 1008 (policy violation) models/gemini-live-2.5-flash-native-audio is not found for API version v1alpha, or is not supported for bidiGenerateContent

Is there any issue with Google AI Studio API ?? Do I need to migrate to Google Vertex AI Studio?

[ravi-incred]

Update, I also tried running the same via @google/genai Javascript library. Getting the same error there as well. Sharing the code snippet used for running the code

           import {
              GoogleGenAI,
              LiveServerMessage,
              MediaResolution,
              Modality,
              Session,
          } from '@google/genai';

            ...code...
      
             const ai = new GoogleGenAI({
                 apiKey: this.apiKey,
             });

            ...code...

            this.session = await ai.live.connect({
                model: this.model,
                callbacks: {
                    onopen: () => {
                        console.log('[GeminiLive] ✅ WebSocket ONOPEN callback triggered!');
                        this.isActive = true;
                        console.log('[GeminiLive] 🟢 Session is now ACTIVE');
                    },
                    onmessage: (message: LiveServerMessage) => {
                        this.responseQueue.push(message);
                        this.handleMessage(message);
                    },
                    onerror: (e: ErrorEvent) => {
                        console.error('[GeminiLive] ❌ WebSocket ERROR callback triggered!');
                        console.error('[GeminiLive] Error details:', e);
                        console.error('[GeminiLive] Error message:', e.message);
                    },
                    onclose: (e: CloseEvent) => {
                        console.log('[GeminiLive] 🔌 WebSocket ONCLOSE callback triggered!');
                        console.log('[GeminiLive] Close reason:', e.reason, 'code:', e.code);
                        this.isActive = false;
                        console.log('[GeminiLive] 🔴 Session is now INACTIVE');
                    },
                },
                config
            });

Error message

[GeminiLive] Close reason: models/gemini-live-2.5-flash-native-audio is not found for API version v1beta, or is not supported for bidiGenerateContent. code: 1008

[bijucyborg]

Same issue for me. The model suddenly stopped working. No changes were made to the code and yet this behaviour suddenly appeared.

{
insertId: "69484941000d06204d72a3ae"
labels: {1}
logName: "projects/xxxxxx/logs/run.googleapis.com%2Fstderr"
payload: "textPayload"
receiveLocation: "us-central1"
receiveTimestamp: "2025-12-21T19:23:46.076984667Z"
resource: {2}
severity: "DEFAULT"
textPayload: "ERROR:google_adk.google.adk.flows.llm_flows.base_llm_flow:Connection closed: received 1008 (policy violation) Publisher Model projects/xxxxxx/locations/global/publishers/google/models/gemini-live-2.5-flash-native-audio wa; then sent 1008 (policy violation) Publisher Model projects/xxxxxxxxx/locations/global/publishers/google/models/gemini-live-2.5-flash-native-audio wa."
timestamp: "2025-12-21T19:23:45.853536Z"
traceSampled: false
}

[vfa-tanna]

+1

[llalitkumarrr]

Hello,

Based on the information shared by @ravi-incred , it appears this issue might be related to the Gemini API rather than the Google ADK.
We recommend running additional tests using google.genai and Vertex AI to isolate the problem:

1. If these tests work correctly: We will investigate the ADK in more detail to identify the cause.
2. If the issue persists with these tools: The problem likely lies with the API. In that case, we recommend reaching out to the dedicated support teams via the google.genai Python/JS GitHub repositories or the Google AI Forum."

[ravi-incred]

@llalitkumarrr after I switched to Vertex AI studio, now it is working

[llalitkumarrr]

Hello @ravi-incred ,

Thank you for the update. Could you please clarify if you used google.genai with VertexAI or google-adk with VertexAI to test this?

If you used google-adk with VertexAI, it implies the issue may be related to the Gemini API rather than the google-adk library itself. If you are comfortable working with VertexAI, you can proceed as is. However, if you prefer not to, we recommend raising this concern on the google.genai GitHub repository or the Google AI forum for faster support."

[Adhishtanaka]

@llalitkumarrr after I switched to Vertex AI studio, now it is working

does it still working?without error.

[ravi-incred]

Yes still working after switching to VERTEX AI

…

On Fri, 9 Jan, 2026, 10:25 am Adhishtanaka Kulasooriya, < ***@***.***> wrote: *Adhishtanaka* left a comment (google/adk-python#3964) <#3964 (comment)> @llalitkumarrr <https://github.com/llalitkumarrr> after I switched to Vertex AI studio, now it is working does it still working?without error. — Reply to this email directly, view it on GitHub <#3964 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/BTX4YIIZSOOP6XPLQLRH7334F4YDTAVCNFSM6AAAAACPMU7FJWVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZTOMRXGE4TEMBUHA> . You are receiving this because you were mentioned.Message ID: ***@***.***>

[d4rkNinja]

Guys, I think this error is due to model deprecations. I encountered the same issue, and after changing the model to the recommended model, it started working.

In Vertex AI, legacy models are still supported, which is why it works there.

You can refer to this link for more details:
https://ai.google.dev/gemini-api/docs/deprecations