Model Evaluation Breakdown

Decompose comprehensive evaluation of {{model}} into subtasks.

Break down into:
1. Accuracy evaluation subtasks
   - Task types to test
   - Metrics per task type
2. Safety evaluation subtasks
   - Risk categories to probe
   - Test case requirements
3. Performance evaluation subtasks
   - Latency tests
   - Throughput tests
4. Integration subtasks
   - API compatibility
   - Error handling

For each subtask, specify inputs, outputs, and dependencies. Create execution schedule for {{timeline}}.