feat(google-gemini/gemini-robotics-er-2-preview): add new models [bot]

[models-bot]

Auto-generated by model-addition-agent for google-gemini/gemini-robotics-er-2-preview.

---

Note

Low Risk
Metadata-only addition with no runtime or auth changes; low risk unless pricing or capability flags are incorrect.

Overview
Adds a new provider catalog file for gemini-robotics-er-2-preview, registering the Google Gemini robotics ER preview model for chat routing and pricing.

The entry marks the model as preview with thinking enabled, serverless provisioning, and multimodal text/image/video/audio inputs. It documents token limits (131K context, 65K max output), sampling params, capabilities (function calling, structured output, prompt caching, code execution, etc.), and per-token/batch/cache costs with links to Google’s robotics docs as sources.

^{Reviewed by Cursor Bugbot for commit 6bc3675. Bugbot is set up for automated code reviews on this repo. Configure here.}

[github-actions]

/test-models

[harshiv-26]

Gateway test results

• Total: 20
• Passed: 20
• Failed: 0
• Validation failed: 0
• Errored: 0
• Skipped: 0
• Success rate: 100.0%

Provider	Model	Scenarios
google-gemini	gemini-robotics-er-2-preview	success: tool-call, tool-call:google-genai, tool-call:stream:google-genai, tool-call:stream, params:stream, params, params:stream:google-genai, params:google-genai, json-output, structured-output, structured-output:stream, structured-output:google-genai, json-output:stream:google-genai, json-output:stream, json-output:google-genai, structured-output:stream:google-genai, reasoning:stream, reasoning, reasoning:stream:google-genai, reasoning:google-genai

Successes (20)

google-gemini/gemini-robotics-er-2-preview — tool-call (success)

Output

Function: get_weather
Arguments: {"location":"London"}
VALIDATION: tool-call SUCCESS

google-gemini/gemini-robotics-er-2-preview — tool-call:google-genai (success)

Output

Tool: get_weather
Args: {'location': 'London'}
VALIDATION: tool-call SUCCESS

google-gemini/gemini-robotics-er-2-preview — tool-call:stream:google-genai (success)

Output

Tool: get_weather
Args: {'location': 'London'}

VALIDATION: tool-call stream SUCCESS

google-gemini/gemini-robotics-er-2-preview — tool-call:stream (success)

Output

{"location":"London"}
VALIDATION: tool-call stream SUCCESS

google-gemini/gemini-robotics-er-2-preview — params:stream (success)

Output

The capital of France is Paris.

google-gemini/gemini-robotics-er-2-preview — params (success)

Output

The capital of France is Paris.

google-gemini/gemini-robotics-er-2-preview — params:stream:google-genai (success)

Output

The capital of France is Paris.

google-gemini/gemini-robotics-er-2-preview — params:google-genai (success)

Output

The capital of France is Paris.

google-gemini/gemini-robotics-er-2-preview — json-output (success)

Output

[
  {
    "name": "Red",
    "hex": "#FF0000"
  },
  {
    "name": "Green",
    "hex": "#00FF00"
  },
  {
    "name": "Blue",
    "hex": "#0000FF"
  }
... (truncated, 36 chars omitted)

google-gemini/gemini-robotics-er-2-preview — structured-output (success)

Output

{
  "name": "science fair",
  "date": "Friday",
  "participants": [
    "Alice",
    "Bob"
  ]
}
VALIDATION: structured-output SUCCESS

google-gemini/gemini-robotics-er-2-preview — structured-output:stream (success)

Output

{"name":"science fair","date":"Friday","participants":["Alice","Bob"]}
VALIDATION: structured-output stream SUCCESS

google-gemini/gemini-robotics-er-2-preview — structured-output:google-genai (success)

Output

{"name":"Science Fair","date":"Friday","participants":["Alice","Bob"]}
VALIDATION: structured-output SUCCESS

google-gemini/gemini-robotics-er-2-preview — json-output:stream:google-genai (success)

Output

[
  {
    "color": "Red",
    "hex": "#FF0000"
  },
  {
    "color": "Green",
    "hex": "#00FF00"
  },
  {
    "color": "Blue",
    "hex": "#0000FF"

... (truncated, 45 chars omitted)

google-gemini/gemini-robotics-er-2-preview — json-output:stream (success)

Output

[
  {
    "name": "Red",
    "hex": "#FF0000"
  },
  {
    "name": "Green",
    "hex": "#00FF00"
  },
  {
    "name": "Blue",
    "hex": "#0000FF"
  }
... (truncated, 43 chars omitted)

google-gemini/gemini-robotics-er-2-preview — json-output:google-genai (success)

Output

[
  {
    "color": "Red",
    "hex": "#FF0000"
  },
  {
    "color": "Green",
    "hex": "#00FF00"
  },
  {
    "color": "Blue",
    "hex": "#0000FF"

... (truncated, 38 chars omitted)

google-gemini/gemini-robotics-er-2-preview — structured-output:stream:google-genai (success)

Output

{"name":"science fair","date":"Friday","participants":["Alice","Bob"]}
VALIDATION: structured-output stream SUCCESS

google-gemini/gemini-robotics-er-2-preview — reasoning:stream (success)

Output

To evaluate the expression $3^{3^{3^3}}$, we must follow the mathematical convention for towers of exponents (tetration). Exponentiation is right-asso
... (truncated, 1472 chars omitted)

google-gemini/gemini-robotics-er-2-preview — reasoning (success)

Output

To calculate the expression $3^{3^{3^3}}$, we must follow the mathematical convention for towers of exponents (tetration). Exponentiation is right-ass
... (truncated, 1453 chars omitted)

google-gemini/gemini-robotics-er-2-preview — reasoning:stream:google-genai (success)

Output

[Thinking] **Analyzing the Expression**

I'm focusing on the expression's structure: $3^{3^{3^3}}$. The right-associativity of exponentiation clarifie
... (truncated, 2311 chars omitted)

google-gemini/gemini-robotics-er-2-preview — reasoning:google-genai (success)

Output

[Thinking] **Calculating a Power Tower: An Exercise in Scale**

Alright, let's tackle this problem. I need to figure out the value of $3^{3^{3^3}}$.  
... (truncated, 3860 chars omitted)

[cursor]

Cursor Bugbot has reviewed your changes using default effort and found 1 potential issue.

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}

^{Reviewed by Cursor Bugbot for commit 6bc3675. Configure here.}

[cursor]

Missing max tokens param override

Medium Severity

params omits a max_tokens override, so the provider default maxValue of 4096 from default.yaml still applies. That conflicts with limits.max_output_tokens / limits.max_tokens of 65536, and with the prior gemini-robotics-er-1.6-preview config that raises max_tokens to 65536. Callers can be capped well below the model’s actual output limit.

 

^{Reviewed by Cursor Bugbot for commit 6bc3675. Configure here.}