voidfeed_benchmark_lookup
Look up model benchmark scores with confidence intervals for 247 model-benchmark combinations.
Instructions
Look up model benchmark scores. Surface tier: 8 entries. The Void tier: 247 model×benchmark combinations with confidence intervals.
Input Schema
TableJSON Schema
| Name | Required | Description | Default |
|---|---|---|---|
| model | No | Model name to look up | |
| benchmark | No | Benchmark name (e.g. MMLU, HumanEval, MATH) |
Implementation Reference
- index.js:129-141 (schema)Schema definition for voidfeed_benchmark_lookup tool: accepts optional 'model' and 'benchmark' string parameters.
{ name: 'voidfeed_benchmark_lookup', description: 'Look up model benchmark scores. Surface tier: 8 entries. The Void tier: 247 model×benchmark combinations with confidence intervals.', inputSchema: { type: 'object', properties: { model: { type: 'string', description: 'Model name to look up' }, benchmark: { type: 'string', description: 'Benchmark name (e.g. MMLU, HumanEval, MATH)' }, }, required: [], }, }, - index.js:215-220 (handler)Handler for voidfeed_benchmark_lookup: builds query params from optional 'model' and 'benchmark' args and calls GET /v1/tools/benchmark-lookup on the VoidFeed API.
case 'voidfeed_benchmark_lookup': { const params = new URLSearchParams(); if (args.model) params.set('model', args.model); if (args.benchmark) params.set('benchmark', args.benchmark); return vfGet(`/v1/tools/benchmark-lookup?${params}`); } - index.js:258-276 (registration)Registration of the tool handler via MCP CallToolRequestSchema, which dispatches to handleTool() for all tools including voidfeed_benchmark_lookup.
server.setRequestHandler(CallToolRequestSchema, async (request) => { const { name, arguments: args } = request.params; try { const result = await handleTool(name, args || {}); return { content: [ { type: 'text', text: JSON.stringify(result, null, 2), }, ], }; } catch (err) { return { content: [{ type: 'text', text: `Error: ${err.message}` }], isError: true, }; } });