The live economics of artificial intelligence.
AI Token Prices: Compare AI Models, APIs and Costs
Token prices across hundreds of models and dozens of providers, three market indices settled daily, and the reporting that explains why they move.
Model Prices
Every model we track, priced per million tokens, checked against the provider’s own page.
The current flagshipsInOutAlso in pricesSee all pricesMarket Indices
Settled daily: what tokens cost, what GPUs cost, and the gap between them.
OpenAI blended rate · last 40 settled days
- TTPI$2.36OpenAI · per 1M tokens, blended-7.60%
- TCPI$3.14H100 SXM · per GPU-hour0.00%
- Frontier Premium×1.0API vs self-hosted-0.89%
Latest settled day: 2026-09-25
MethodologySee the indicesCost Calculator
Your own token volumes, priced model by model.
20M4M$220.00 a month
on GPT-5.6 Sol
The same month on Gemini 3.8 Flash: $30.00
Drag either volume. Full calculator counts your real tokens.
Compare
Four models side by side, on price, context and benchmark score.
GPT-5.6 Sol costs 7.7× the blended rate of Grok 4.3.
Rankings
Ranked lists computed from live prices: cheapest, best value, longest context.
Cheapest AI ModelsBenchmarks
Leaderboards that carry the price beside every score, not just the score.
SWE-benchScoreBlendedOther leaderboardsAll leaderboardsGuides & Glossary
Plain-language explainers on how AI pricing works, and what the words mean.
From the glossaryContext windowThe context window is the maximum number of tokens a model can handle in a single request, covering the input prompt and generated output combined.
The signal behind the numbers.
MODELSOpenAI pauses training of its ‘most capable models’As reports of OpenAI's models breaking containment, hacking sites, and generally getting out of control pile up, the company has made the decision to pause training of its most powerful models. The decision was made after a model being tested within a sandbox exploited a loophole to gain internet access. The incident happened on September 20th, and "All training, evaluation, and inference with tool-use" remains paused as of Saturday evening, September 25th. In addition, OpenAI revealed on…The Verge AI3h ago- OpenAI expands review of model behavior after more rogue agent incidents emergeCNBC Technology2h ago
- Casar wants Democrats to make JD Vance’s Silicon Valley ties a liabilityPolitico Tech3h ago
- The blue-collar AI job market is booming. Will data center backlash make it go bust?CNBC Technology5h ago
- Can Cloudflare CEO Matthew Prince save the web from AI?The Verge AI6h ago
Big Reads
- 02Ahead of US IPO, British AI neocloud Nscale secures $3.36B in convertible financing
- 03How OpenAI’s Rogue A.I. Agents Tried to Trick a Robot Detector
- 04Oxford lets OpenAI train its AI models on Bodleian Library
- 05OpenAI says agents leaked 53 images from ChatGPT users in latest example of rogue activity
- 06Meta’s Muse just stole the AI spotlight from OpenAI and Anthropic
- 07Accel Sees AI Opportunity Shifting to Applications
- 08Anthropic Goes Big on Compute, Microsoft Rethinks AI
- 09OpenAI expands review of model behavior after more rogue agent incidents emerge
- 10Proaction boosts sales 60% and saves 75+ hours with Codex
BY CATEGORY
Across the AI economy.
Frequently asked questions
What is Tokenando?
Tokenando tracks the live economics of AI: token prices across hundreds of AI models from dozens of providers, three market indices settled daily, head-to-head model comparisons, and the reporting that explains why prices move.
Where does Tokenando get its pricing data?
Prices are aggregated continuously from live marketplace endpoints, including OpenRouter, and hand-curated from official provider pricing documentation. We never estimate, interpolate, or infer a price: a model with no published price is listed without one. Large price movements are held in a review queue rather than published automatically.
Which AI providers and models are tracked?
OpenAI, Anthropic, Google, Meta, Mistral, DeepSeek, xAI, Cohere, Moonshot, Amazon Bedrock, Azure, Together, Fireworks, Groq, OpenRouter and dozens more. New providers and models are added as they launch, and retired models stay readable for 90 days before they are removed.
How often are AI token prices updated?
The pricing index refreshes daily and the site rebuilds after it, so every page serves prices read within the last 24 hours. The news feed reads 25 active publisher sources and refreshes through the day.
What is a blended token price?
Blended cost is our own single-number summary, weighting the input rate at 70 per cent and the output rate at 30 per cent so models with different input and output prices can be ranked side by side. It is a calculation of ours, not a rate any provider quotes, and we label it that way wherever it appears.
Is Tokenando free?
Yes. The price index, the calculator, the comparison tools, the rankings, the benchmark leaderboards, the guides and the news feed are all open to anyone, with no account required.
Track the economics of AI before they move against you.
Explore live model pricing, compare provider economics, and follow the market forces shaping the cost of intelligence.
CONTACT
Talk to Tokenando
Press, data corrections, partnership ideas, or feedback on the platform, the right person on the team will reply directly. We read every message.