| 25 | } |
| 26 | |
| 27 | export interface BenchmarkEntry { |
| 28 | provider: string; |
| 29 | family: 'claude' | 'gpt' | 'gemini'; |
| 30 | available: boolean; |
| 31 | unavailable_reason?: string; |
| 32 | result?: RunResult; |
| 33 | costUsd?: number; |
| 34 | /** Judge score 0-10 across dimensions. Populated separately by the judge step. */ |
| 35 | qualityScore?: number; |
| 36 | qualityDetails?: Record<string, number>; |
| 37 | } |
| 38 | |
| 39 | export interface BenchmarkReport { |
| 40 | prompt: string; |
nothing calls this directly
no outgoing calls
no test coverage detected