LLMs/AIX: Gemini: computer use test

This commit is contained in:
Enrico Ros
2025-11-01 11:18:12 -07:00
parent 6a792814ce
commit 628b88ef9f
2 changed files with 19 additions and 3 deletions
+2 -1
View File
@@ -159,7 +159,8 @@ export const DModelParameterRegistry = {
type: 'enum' as const,
description: 'Environment type for Computer Use tool (required for Computer Use model)',
values: ['browser'] as const,
requiredFallback: 'browser',
initialValue: 'browser',
// requiredFallback: 'browser', // See `const _requiredParamId: DModelParameterId[]` in llms.parameters.ts for why custom params don't have required values at AIX invocation...
hidden: true,
} as const,
@@ -44,7 +44,7 @@ const filterLyingModelNames: GeminiWire_API_Models_List.Model['name'][] = [
- Latest stable version gemini-1.0-pro <model>-<generation>-<variation>
- Stable versions gemini-1.0-pro-001 <model>-<generation>-<variation>-<version>
Gemini capabilities chart (updated 2025-09-29):
Gemini capabilities chart (updated 2025-11-01):
- [table stakes] System instructions
- JSON Mode, with optional JSON Schema
- Adjustable Safety Settings
@@ -65,7 +65,7 @@ const geminiExpFree: ModelDescriptionSchema['chatPrice'] = {
};
// Pricing based on https://ai.google.dev/pricing (Sept 29, 2025)
// Pricing based on https://ai.google.dev/pricing (Nov 1, 2025)
const gemini25ProPricing: ModelDescriptionSchema['chatPrice'] = {
input: [{ upTo: 200000, price: 1.25 }, { upTo: null, price: 2.50 }],
@@ -239,6 +239,21 @@ const _knownGeminiModels: ({
benchmark: { cbaElo: 1424 },
},
// 2.5 Pro-Based: Gemini Computer Use Preview - Released October 7, 2025
{
id: 'models/gemini-2.5-computer-use-preview-10-2025',
labelOverride: 'Gemini 2.5 Computer Use Preview 10-2025',
isPreview: true,
chatPrice: gemini25ProPricing, // Uses same pricing as 2.5 Pro (pricing page doesn't list separately)
interfaces: [LLM_IF_OAI_Chat, LLM_IF_OAI_Vision, LLM_IF_OAI_Fn, LLM_IF_OAI_Json, LLM_IF_OAI_Reasoning, LLM_IF_GEM_CodeExecution],
parameterSpecs: [
{ paramId: 'llmVndGeminiThinkingBudget' },
{ paramId: 'llmVndGeminiComputerUse' },
],
benchmark: undefined, // Computer use model, not benchmarkable on standard tests
hidden: true, // Specialized browser automation model - show when needed for specific use cases
},
// 2.5 Flash-Based: Gemini Robotics-ER 1.5 Preview - Released September 25, 2025
{
id: 'models/gemini-robotics-er-1.5-preview',