Free LLM & AI API pricing comparison
Compare vendor list prices side by side, pick a workflow below, then open the full table—or jump to a monthly cost estimate.
Choose what you are optimizing for
These presets only change filters and sorting—the full catalog stays available below.
Tokens are how APIs meter usage. We quote USD per 1 million tokens so you can scan prices without tiny decimals. Input is what you send; output is what the model writes back (often pricier). See column definitions or the short methodology.
Prefer to browse raw SKUs? Jump to the comparison table, export CSV, or open the compare drawer (up to three models).
- List-price estimates — not invoices
- Confirm on vendor consoles for quotes
- How we normalize this data
- Source: aggregated public catalog
Popular picks for chat-style workloads
Five models with the lowest listed output price among text-capable SKUs in this snapshot—sorted for scanning, not a quality ranking.
Explore every model
Search, filter, compare up to three SKUs side by side, group by provider, or download CSV.
Providers
Filters and sorting
More filters and sorting
What do these columns mean?
- Input cost — What you pay to send your prompt (questions, documents, code).
- Cached input — Discounted rate when providers reuse repeated prompt parts.
- Output cost — What you pay for the model’s reply (usually higher than input).
- Context window — How much text fits in one conversation at once (bigger = longer memory).
- Max response — Cap on how long a single reply can be, when listed.
| Compare | Provider | Model id | Display name | Input cost / 1M tokens | Cached input / 1M | Output cost / 1M tokens | Context window | Max response | Modality | Catalog |
|---|---|---|---|---|---|---|---|---|---|---|
Why use this free LLM API pricing tool?
People search for OpenAI API pricing, Claude API cost, Gemini API price per million tokens, and the cheapest LLM for chat—usually before a roadmap or budget review. CloudyBot’s comparison is a no-signup, free browser tool: a live sortable table of published list rates in USD per 1M tokens, a free-tier filter, quick picks for common workflows, and a three-model compare drawer. Use it to shortlist SKUs, then confirm live pricing on each vendor’s billing console.
- LLM price comparison across vendors — see input, cached (where listed), and output $/1M in one grid.
- “Free” list-tier models — filter rows where the catalog shows $0 list input and output (limits may still apply).
- Next steps — estimate rough monthly spend with the free AI cost calculator; run agents on CloudyBot plans when you want hosted automation, not just spreadsheets.
What is AI API pricing per 1M tokens?
Tokens are the billing units most LLM APIs expose in pricing consoles. Showing dollars per million tokens lets you compare SKUs without mentally juggling scientific notation from per-token rates.
Figures here come from a normalized catalog snapshot (aggregated public model listings, reconciled to USD per 1M tokens on our servers). Treat them as directional list estimates—not a contractual quote—until you confirm live rates in each vendor’s billing console.
Input vs output pricing (quick example)
If input is $2 / 1M tokens and output is $8 / 1M, a prompt with 800 input tokens and 400 output tokens would cost roughly $0.0048 — always verify rounding and minimum charges on the vendor side.
Cheapest models by workload shape
Chat-heavy assistants usually care most about average output price. RAG or tool-heavy flows may spike input tokens. Vision SKUs add image-token lanes—filter modality to narrow candidates.
Provider pricing guides
Jump to focused tables with the same snapshot normalization:
Official vendor billing pages
For invoice-critical quotes, cross-check each vendor console:
Recent catalog updates
Published rows refresh when operators ship a validated snapshot. Large vendor repricing may queue review before replacing the live table—reload after announcements if numbers look stale.
FAQ
How do I compare OpenAI, Anthropic, and Google AI API pricing in one place?
Use the sortable table: search or filter by provider, then read input and output USD per 1M tokens side by side (OpenAI-style GPT, Anthropic Claude, Google Gemini, and more). For quotes, still confirm on each vendor’s billing page.
Is this LLM pricing comparison free?
Yes. Browsing, filters, three-way compare, and CSV export do not require a CloudyBot account. Numbers are list estimates from a catalog snapshot, not your invoice.
How do I find the cheapest LLM API for chatbots?
For assistants that mostly generate replies, sort by output price per 1M on text-capable models—then validate quality and latency on your own prompts before you lock in.
Are model IDs OpenRouter-style?
Rows use normalized ids from an aggregated public catalog (often similar to OpenRouter-style model keys). Always verify the exact SKU and rate on your router or provider console.
What does USD per 1M tokens mean?
Estimated dollars for one million billing tokens at published list rates.
Input vs output?
Input covers prompts; output covers generations—often priced higher.
Cached input?
Discounted reads for repeated prefixes where vendors expose separate lanes.
Where does data come from?
A public catalog normalized server-side; confirm on vendor billing consoles.
How often is it updated?
On a scheduled cadence plus manual publishes after validation—not real-time per keystroke.
Pick the cheapest LLM API?
Sort by the price column that dominates your trace, filter modality, pilot quality.
Free models?
Filtered rows show SKUs where catalog list input/output are both zero—limits may still apply.
What is context window?
Maximum tokens held in a single request window for that SKU.
What is a token?
A billing slice of your text—providers charge per token; we show dollars per million tokens so numbers stay readable.
Cheapest ChatGPT-style API?
List rates move often—sort by output cost for chat-heavy apps, confirm on the vendor console before you budget.
Free AI APIs?
Some rows are $0 list price; quotas and limits still apply—check each vendor.
Monthly spend estimates?
Use the AI cost calculator after picking rates here.
Automate after you pick a model
CloudyBot is an AI worker platform—chat, browser automation, and workflows—not just a pricing table. See how it works, compare plans, or open the dashboard.