By name: model-advisor description: Recommends the optimal Claude model based on task requirements and cost analysis type: semantic user-invocable: true semantic: skillVersion: 1 permissions: graph: [read] memory: [recall] queries: - predicate: "advisor:recommendation" scope: agent
model-advisor
Recommends the optimal Claude model based on task requirements, complexity analysis, and cost efficiency.
When to Use
Use when choosing between Claude models for a specific task, when optimizing API costs, or when you need guidance on which model best fits a workload profile.
Supported Models
| Model | Input | Output | Best For |
|---|---|---|---|
| Opus 4.6 | $15/MTok | $75/MTok | Complex reasoning, research, multi-step analysis, code architecture, creative writing |
| Sonnet 4.6 | $3/MTok | $15/MTok | Coding, general tasks, data analysis, moderate complexity |
| Haiku 4.5 | $0.80/MTok | $4/MTok | Classification, extraction, simple Q&A, high volume, low latency |
Instructions
- Recall any previous model recommendations with
skill_memory_context - Analyze the task description from query results for complexity indicators
- Match task keywords against model strengths to determine the best fit
- Return the recommended model with rationale and cost comparison