> ## Documentation Index > Fetch the complete documentation index at: https://docs.corepillar.ai/llms.txt > Use this file to discover all available pages before exploring further. # Explain a score gap between two models > Writes one or two sentences explaining why two models scored differently on the same evaluator, from the judge's own notes. Uses a small platform-funded model call against your organization's platform allowance; when that isn't available or fails, returns a factual description of the score difference with `source: "heuristic"`. **Auth:** session. **Permission:** `run_agents` or `manage_agents`. ## OpenAPI ````yaml /api-reference/openapi.json post /projects/{projectId}/evaluations/explain openapi: 3.1.0 info: title: CorePillar API version: 1.0.0 description: >- REST API for AgentHub and ModelHub. Ingestion endpoints authenticate with a project tracing key; all other endpoints use the signed-in user's browser session. servers: - url: https://pillarforge.corepillar.ai/api/v1 description: AgentHub security: [] tags: - name: Ingestion description: 'Send traces with a tracing key: JSON ingest and OpenTelemetry (OTLP).' - name: Traces description: Read traces and spans. - name: Evaluations description: Evaluate traces with evaluator collections. - name: Evaluators description: Built-in and custom evaluators. - name: Evaluator collections description: Ordered groups of evaluators. - name: Organizations description: Organizations you belong to. - name: Members description: Organization members, invites and roles. - name: Projects description: Projects inside an organization. - name: SSO description: Okta SSO configuration. - name: Tracing keys description: Create, list and revoke project tracing keys. - name: Cost calculator description: Estimate, generate and execute scenario runs. - name: Synthetic description: Synthetic agent and data setup from source material. - name: Chat description: ModelHub chat. Served at https://modelhub.corepillar.ai/api/v1. - name: Auth description: Sign-in helpers. - name: E2E description: Internal end-to-end encryption key directory. paths: /projects/{projectId}/evaluations/explain: post: tags: - Evaluations summary: Explain a score gap between two models description: >- Writes one or two sentences explaining why two models scored differently on the same evaluator, from the judge's own notes. Uses a small platform-funded model call against your organization's platform allowance; when that isn't available or fails, returns a factual description of the score difference with `source: "heuristic"`. **Auth:** session. **Permission:** `run_agents` or `manage_agents`. operationId: explainScoreGap parameters: - name: projectId in: path required: true description: Project id. schema: type: string format: uuid requestBody: required: true content: application/json: schema: type: object required: - evaluator_name - model_a - model_b properties: evaluator_name: type: string minLength: 1 maxLength: 160 model_a: $ref: '#/components/schemas/ExplainSide' model_b: $ref: '#/components/schemas/ExplainSide' responses: '200': description: Explanation. content: application/json: schema: type: object required: - data properties: data: type: object properties: explanation: type: string maxLength: 400 source: type: string enum: - model - heuristic '400': description: Validation failed. `error.code` is `VALIDATION_ERROR`. content: application/json: schema: $ref: '#/components/schemas/ErrorEnvelope' '401': description: Not signed in (`UNAUTHORIZED`). content: application/json: schema: $ref: '#/components/schemas/ErrorEnvelope' '403': description: >- Not a member of the organization, or missing the required permission (`FORBIDDEN`). content: application/json: schema: $ref: '#/components/schemas/ErrorEnvelope' '404': description: Project not found or not visible to you (`PROJECT_NOT_FOUND`). content: application/json: schema: $ref: '#/components/schemas/ErrorEnvelope' '500': description: Unexpected server error (`INTERNAL_ERROR`). content: application/json: schema: $ref: '#/components/schemas/ErrorEnvelope' security: - session: [] components: schemas: ExplainSide: type: object required: - label - score - passed properties: label: type: string minLength: 1 maxLength: 120 score: type: number minimum: 0 maximum: 100 passed: type: boolean reasoning: type: string maxLength: 2000 default: '' criteria: type: array maxItems: 8 default: [] items: type: object required: - name - score properties: name: type: string minLength: 1 maxLength: 64 score: type: number minimum: 0 maximum: 1 note: type: string maxLength: 240 default: '' issues: type: array maxItems: 6 default: [] items: type: object required: - category - severity properties: category: type: string minLength: 1 maxLength: 64 severity: type: string enum: - none - low - medium - high - critical detail: type: string maxLength: 240 default: '' evidence: type: array maxItems: 4 default: [] items: type: string maxLength: 240 ErrorEnvelope: type: object required: - error properties: error: type: object required: - code - message properties: code: type: string description: Stable machine-readable code. Branch on this, not on `message`. examples: - VALIDATION_ERROR message: type: string description: Human-readable description. May change. details: type: object additionalProperties: true description: Extra context; often empty. securitySchemes: session: type: apiKey in: cookie name: session description: >- The session cookie set when you sign in to AgentHub or ModelHub in a browser. The cookie name shown here is illustrative. There are no API tokens for these endpoints. ```` This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.