There are 350+ AI models across 40+ providers — and their prices, speed and availability change every week. Pick the wrong one and you can pay 5–20× more for the same quality, or wake up to an app that broke because a model was quietly retired. The LLM Price Compass is the free, independent dashboard that shows you — in plain language — the best model for the job and what it really costs, updated live every hour.
AI got cheap and powerful fast — but the market is now so big and moves so quickly that almost everyone is leaving money on the table without realizing it.
The cheapest capable model is often 10–50× cheaper than the popular default. Most people pick one model on day one and never look again — and the bill quietly compounds, month after month.
New models launch, prices drop, and old ones get deprecated constantly. What was the smart choice last month can be twice the price — or simply gone — today.
350+ models, dozens of providers, separate input vs output prices, hidden latency. No human can keep this straight. So we do it for you — automatically, every hour.
If anything you do touches an AI API, this is for you — whether you write code or not.
You use AI but don't have an engineer watching the bill. See exactly what you're spending and where to cut it — no code, no jargon.
Pick the right model per task, route to the cheapest host, and pull our live data straight into your app or dashboard through a free API.
Your agents burn tokens 24/7. Keep them on the best-value model and get alerted the moment one is deprecated or repriced — before it breaks something.
Turn an opaque "AI" line item into a clear, defensible number — with a concrete switch-and-save plan attached.
Four tools, one dashboard — the full picture of cost, change, spend and uptime across the AI market.
Live side-by-side prices, latency and context window for every major model. Filter, sort by cheapest, and copy a ready-to-run config in one click.
New models, price drops and deprecations the moment they happen — so nothing silently breaks your app or quietly wastes your money.
Connect your provider account and see your real spend, broken down by model — plus exactly which switches would save you the most.
Live up/down and latency for every major AI API, measured from real servers, with uptime history so you know who to trust.
Two ways — pick yours. No account required either way.
Head to the homepage. Every model is listed with its real price, speed and context window — all live.
Tier 1 for frontier quality, Tier 2 for value, "100% Free" for zero-cost models, or filter by latency if speed matters. The best-value option floats to the top.
Hit "Copy config" on any model and paste it into your tool or codebase. That's it — you're now running on the best-value model for the job.
Watch the Change Radar (or get alerts on a paid plan) so you always know the moment something cheaper, faster or better lands — and before anything you rely on gets retired.
Prefer code? Jump to the developer API below.
Other "comparisons" are blog posts that are out of date the day they publish, or dev tools that need an engineering team to set up. This is neither.
Refreshed every hour from the source — never a quarterly article that's already wrong.
Tier 1 and Tier 2, every major provider and 350+ models in one view — not just the big three.
We don't sell model inference, so we've no reason to push you toward the expensive option.
We probe providers from real servers — "fast" means fast for you, not a number on a spec sheet.
The full comparison, the Change Radar and a free API tier, with no account required.
A dashboard a founder can read, and an API your code or your agents can call.
Everything on this site is also a free, CORS-enabled JSON API. Build it into your app, your agent, or an internal dashboard — no key needed to start.
import requests
r = requests.get("https://api.dunamai.media/v1/models",
params={"tier": 1, "max_latency": 800})
m = sorted(r.json()["models"], key=lambda x: x["blended_per_mtok"] or 9e9)[0]
print(m["id"], "→ $", m["blended_per_mtok"], "/Mtok")
curl "https://api.dunamai.media/v1/cheapest?model=llama-3.1-70b"
Higher rate limits plus premium data — Change Radar webhooks, price history and cheapest-host routing — come with an API key. See the plans →
No signup, no code. Compare every model live, or grab the free "State of LLM Pricing" report.