AI pricing explorer

AI model pricing: compare LLM API prices by provider

Filter by workload, refine the latest snapshot, pin rows to compare, and open inline history without leaving the table.

Filter rail

Compare current and earlier models by workload, provider, and verified lifecycle status.

All models
Type or choose providers
Type or choose routes
Type or choose categories
Type or choose workloads
Type or choose capabilities
Choose model statuses
Choose sorting
Default

Results

Loading…

Token rates per 1M tokens. — means no listed rate.

Useful for shortlist validation

  • The planner recommends. This explorer lets you audit the rows, route categories, sync freshness, and recent changes behind those recommendations.
Useful for research
Workload chips open common questions quickly, while the table remains flexible for deeper provider, source, and billing-unit analysis.

What this page is for

Start from workload chips
Jump into chat, embeddings, document, voice, image, video, search, or long-context rows before narrowing the results.
Pin and compare
Pin up to four rows in a compact comparison tray without losing the current filtered view.
Open row history
Each row has a sparkline and expandable history so you can see whether a price is new, stable, or recently changed.

Useful with the main article

Need market context and written analysis too? Read the full 2026 AI pricing comparison.

Frequently asked questions

Where does the pricing data come from?

Rows come from the site's pricing database and retain their provider or aggregator source. Use the source link on a row to verify the current public price.

Can I compare different billing units?

Yes, but treat token, request, image, minute, and other unit prices as different dimensions. The explorer shows each unit rather than pretending every row is directly equivalent.

Are filters and pinned rows shareable?

Yes. The current search, filters, sorting, and pinned row identifiers stay in the URL when you copy the current view.