Free tool

# LLM Selector: Which AI Model Should I Use?

> Answer four questions and get an evidence-backed shortlist of AI models for your job, budget and constraints. Free, no sign-up.
- Canonical page: https://noometry.com/tools/llm-selector
- Last updated: 2026-10-10
- Title: LLM Selector: Which AI Model Should I Use? | Noometry

Pick the job, your budget and any hard constraints. The selector blends the benchmark categories that matter for that job and returns a shortlist with the reasons behind each pick.

Last verified October 10, 2026

What is the job? 

Max blended price ($ per M tokens) 

Minimum context window 

Priority 

 Open weights only

1.  [**GPT-6 Astra**](https://noometry.com/models/gpt-6-astra)  
    Everyday assistant fit 77.5, #1 overall, $20.00 per M tokens blended
2.  [**Claude Fable 5.1**](https://noometry.com/models/claude-fable-5-1)  
    Everyday assistant fit 75.8, #2 overall, $20.00 per M tokens blended
3.  [**Claude Opus 5.5**](https://noometry.com/models/claude-opus-5-5)  
    Everyday assistant fit 75.4, #3 overall, $8.00 per M tokens blended

## Current top picks by job

### Everyday assistant

1.  [GPT-6 Astra](https://noometry.com/models/gpt-6-astra): Everyday assistant fit 77.5, #1 overall, $20.00 per M tokens blended
2.  [Claude Fable 5.1](https://noometry.com/models/claude-fable-5-1): Everyday assistant fit 75.8, #2 overall, $20.00 per M tokens blended
3.  [Claude Opus 5.5](https://noometry.com/models/claude-opus-5-5): Everyday assistant fit 75.4, #3 overall, $8.00 per M tokens blended

### Coding

1.  [Claude Fable 5.1](https://noometry.com/models/claude-fable-5-1): Coding fit 69.4, #2 overall, $20.00 per M tokens blended
2.  [GPT-6 Astra](https://noometry.com/models/gpt-6-astra): Coding fit 68.9, #1 overall, $20.00 per M tokens blended
3.  [Claude Fable 5](https://noometry.com/models/claude-fable-5): Coding fit 67.6, #5 overall, $20.00 per M tokens blended

### Agents and tool use

1.  [Claude Opus 5](https://noometry.com/models/claude-opus-5): Agents and tool use fit 64.6, #4 overall, $10.00 per M tokens blended
2.  [GPT-6 Astra](https://noometry.com/models/gpt-6-astra): Agents and tool use fit 64.0, #1 overall, $20.00 per M tokens blended
3.  [Claude Fable 5](https://noometry.com/models/claude-fable-5): Agents and tool use fit 63.5, #5 overall, $20.00 per M tokens blended

### Writing

1.  [Claude Fable 5.1](https://noometry.com/models/claude-fable-5-1): Writing fit 79.2, #2 overall, $20.00 per M tokens blended
2.  [Claude Opus 5](https://noometry.com/models/claude-opus-5): Writing fit 79.2, #4 overall, $10.00 per M tokens blended
3.  [Claude Opus 5.5](https://noometry.com/models/claude-opus-5-5): Writing fit 78.7, #3 overall, $8.00 per M tokens blended

## Frequently asked questions

### How does the LLM selector choose models?

It blends the benchmark categories that matter for the job you pick (for example coding weights coding and agentic results), removes models that break your constraints, and optionally trades score for price. Every pick lists the reasons.

### Can I share a result?

Yes. Your answers are stored in the page URL, so copying the link reproduces the same shortlist.

### Is there an API?

Yes: GET /api/recommend?job=coding&budget=5&open=1 returns the same shortlist as JSON.