How do filter stacks work?

Toggle any combination of filters to narrow the model set. All active filters apply together. For example, "Vision + Open source + Reasoning" shows only open-weight models that can see images AND do chain-of-thought reasoning.

What is the Value score?

Value = Intelligence ÷ price. It measures how much capability you get per dollar. Higher is better — a model with IQ 80 at $0.50/M is better value than IQ 85 at $3.00/M.

How do I save a model for later?

Click the star icon next to any model. Your starred models are saved in this browser. Toggle the "Favorites" filter to see only your starred set.

Where does the data come from?

Pricing from the OpenRouter API, updated daily. Intelligence scores from aggregated benchmark rankings, normalized to 0-100.

Find the right AI model for your task

Compare 300+ LLMs across 50+ providers. Tell us what you're building and we'll show you the best options for your budget and use case.

Click a task above to jump to the table with tips

Sort by

Filter

How to use

Star models to build your own set

Sort by IQ for the smartest models

Stack filters to narrow by capability

Quick help

Pricing from OpenRouter · Intelligence scores from public benchmarks · Updated daily

How to choose the right LLM API

Picking the best model for your project depends on three factors: capability, cost, and context. The table above compares 300+ models across 50+ providers so you can find the optimal balance. Use the search bar to find a specific model, stack filters (Vision + Open source) to narrow results, or sort by Intelligence to find the smartest models in your budget.

Cheapest pick

Sort by price ascending. Many capable models cost under $0.50/1M input tokens.

Best value

Sort by Value score — balances intelligence against cost.

Smartest models

Sort by IQ to see the highest-ranked models by benchmark performance.

Large context

Sort by context window — some models handle 1M+ tokens.

Frequently asked questions about LLM pricing

How do I find the cheapest LLM for my use case?▼

Sort the table by Input Price (ascending) to see the most affordable models. Then toggle filters for specific capabilities — Vision, Open source, or Reasoning — to narrow results. For a personalized recommendation, use the Cost Calculator to match your usage volume, token ratio, and priority (cheapest, fastest, smartest).

What does the Value score mean?▼

Value = Intelligence ÷ log(price). It measures how much capability you get per dollar spent. A mid-priced model with strong benchmarks often beats a cheap model with poor performance on value. Sort by Value to find the best bang for your buck.

Should I use a subscription (e.g. OpenCode Go, ChatGPT Plus) or API pricing?▼

It depends on your volume. Subscriptions are better for moderate usage (under ~10M tokens/month) and include free rate limits. API pricing is more cost-effective at high volumes and gives you access to the latest models. Use the Cost Calculator with your actual token counts to compare.

Why do the same models have different prices across providers?▼

Different providers negotiate different rates with model creators, add their own margins, or subsidize certain models to attract users. For example, DeepSeek R1 is available on both OpenRouter and Together AI at different prices. The table shows each provider's listed price — check multiple rows for the same model name.

How often is the pricing data updated?▼

Pricing is fetched from the OpenRouter API and cached for up to 1 hour. Models and prices change frequently — this site reflects the latest available pricing from each provider's API.

Which model has the largest context window?▼

Sort the table by Context (descending) to see models with the biggest context windows. Current leaders include Gemini 2.5 Pro (1M+ tokens), DeepSeek models, and Anthropic Claude Sonnet 4 (200K tokens).

Last updated: May 21, 2026Data source: OpenRouter APIPrivacySitemapCost Calculator