AI AI Prices
Public pricing tracker

Use one site to understand AI API prices, choose the next tool, and trust the number you are seeing.

AI Prices turns public token pricing into a clear first-visit flow: scan the market, compare the same workload, estimate budget impact, or inspect recent pricing changes without rebuilding the logic yourself.

Start with calculator Compare workload See recent changes Browse models
Calculator for monthly spend Compare for fixed workload Changes for market movement Models for broad scanning

Homepage pricing switches to the selected display currency when FX is available. If conversion fails, cards stay in the source currency shown on the amount.

Recommended path

Know token volume first? Open calculator. Need apples-to-apples math? Open compare. Need to understand the market? Open models. Need to see what changed lately? Open changes.

Providers tracked 3 3 with live snapshots
Models indexed 16 Unified token cost catalog
Last updated Apr 12, 22:05 Latest stored snapshot
Source policy Official-first Verified fallback stays labeled when official pages block crawlers.

Best entry prices right now

Sorted by input price in USD when FX is available. Blend = 75% input + 25% output.

See full ranking
Rank Model Input Blend Updated
1 GPT-4.1 Nano
OpenAI
 $0.050 $0.088
Apr 12, 22:05 PricePerToken OpenAI
2 GPT-5 Nano
OpenAI
 $0.050 $0.138
Apr 12, 22:05 PricePerToken OpenAI
3 Gemini 2.5 Flash-Lite
Gemini
 $0.100 $0.175
Mar 30, 02:54 Gemini Developer
4 GPT-5 Mini
OpenAI
 $0.125 $0.344
Apr 12, 22:05 PricePerToken OpenAI
5 GPT-4o Mini
OpenAI
 $0.150 $0.263
Apr 12, 22:05 PricePerToken OpenAI
6 GPT-4.1 Mini
OpenAI
 $0.200 $0.350
Apr 12, 22:05 PricePerToken OpenAI

Start here in 30 seconds

The homepage should answer two things fast: what the site does, and which page you should open next.

Need a budget answer Open calculator

Best when token shape, request volume, or budget already exists and you need monthly spend, cache savings, or budget fit.

Go to calculator
Need a same-workload comparison Open compare

Best when the request shape is fixed and you want one side-by-side answer instead of sorting single price fields.

Go to compare
Need market movement Open changes

Best when you care about fresh up or down signals, not just the latest stored point.

Go to changes
Need broad catalog scan Open models

Best when you are still exploring providers, freshness, cached pricing coverage, and headline token rates.

Go to models

Why teams can trust this pricing

This section exists to reduce trust friction before you make a cost decision or share the numbers with someone else on the team.

Methodology Data sources

Official-first, with verified fallback

Each stored price prefers the provider's public pricing page first. If that path blocks crawlers, the site keeps a verified fallback label visible instead of pretending the number came from a live official crawl.

Timestamped stored snapshots

Every number is tied to a stored snapshot time, so freshness is visible instead of implied by the render time of the page.

Normalized pricing terms

Input, output, cached, and batch fields are translated into the same pricing language across home, list, compare, and calculator views.

Decision tools use the same pricing base

The calculator and compare flows read from the same pricing layer instead of a separate rough estimation table.

Coverage snapshot

A quick operational summary of what is live now, where deeper pricing detail already exists, and where coverage still waits on fresh snapshots.

Live model pricing 16 / 16 Models with a stored pricing snapshot right now.
Providers live now 3 / 3 Providers already feeding ranking, compare, and calculator flows.
Providers awaiting snapshots 0 Catalog exists, but fresh pricing is not stored yet.
Models with listed batch pricing 3 Counts models where a normalized batch price is currently listed, not blanket provider-wide support.

Browse by provider

Provider pages are the fastest way to inspect one ecosystem at a time, including freshness, model count, and current pricing depth.

Open provider index
ANTHROPIC

Anthropic

Open provider

Live snapshot coverage is available for this provider. Official-first sourcing stays preferred, and any verified fallback remains visible on model detail pages.

4 models 4 cached 0 with listed batch pricing

Counted from each model's latest public snapshot where a normalized batch discount field is already present.

0 tracked changes No recent delta yet
Provider detail Provider changes
GEMINI

Gemini

Open provider

Live snapshot coverage is available for this provider. Official-first sourcing stays preferred, and any verified fallback remains visible on model detail pages.

3 models 3 cached 3 with listed batch pricing

Counted from each model's latest public snapshot where a normalized batch discount field is already present.

0 tracked changes No recent delta yet
Provider detail Provider changes
OPENAI

OpenAI

Open provider

Live snapshot coverage is available for this provider. Official-first sourcing stays preferred, and any verified fallback remains visible on model detail pages.

9 models 9 cached 0 with listed batch pricing

Counted from each model's latest public snapshot where a normalized batch discount field is already present.

40 tracked changes Latest change Apr 12, 22:05
Provider detail Provider changes

How to read pricing

Most confusion comes from a few repeated pricing terms. Keep these definitions consistent before you interpret any ranking or calculator result.

Glossary FAQ Methodology Data sources

Input price

The cost for prompt tokens. Prompt-heavy, retrieval-heavy, or tool-heavy workloads should scan this first because it often drives baseline spend.

Output price

The cost for generated tokens. Long-form generation, summarization, and agent responses often lean on this field more heavily.

Cached pricing

A discounted input rate for reusable prompt context when the provider exposes it clearly enough to normalize.

Batch and blend

Blend is a ranking shortcut based on 75% input and 25% output. Batch counts on this site mean a provider-listed batch price could be normalized for that model, not that every provider workflow runs that way by default.

What to do first

The site supports four common first steps. Pick the one that matches the decision you need to make right now.

Step 1 for budget planning Start with the calculator

Use calculator first when you already know rough token volume and need monthly spend, cache leverage, batch savings, or budget fit.

Monthly spend Budget fit Scenario templates
Open calculator
Step 1 for shortlist decisions Compare the same workload

Use compare when the request shape is already known and you want a side-by-side answer without rebuilding the math in a spreadsheet.

Winner cards Same workload
Open compare
Step 1 for fresh signals Check the changes feed

Use changes first when price motion matters more than the current static table and you need to see recent up or down events across providers.

Recent deltas Provider filters
Open changes
Step 1 for market scanning Browse the full model list

Use the model list when you need a wider scan across providers, price signals, freshness, and capability tags before narrowing the shortlist.

Coverage scan Freshness Capability filters
Browse models

Reference workload snapshot

Example scenario: 2k input + 1k output, 20% cache hit, 300k requests / month. Same assumptions, three low-cost models, one faster monthly sanity check.

Adjust assumptions
OpenAI

GPT-4.1 Nano

Run it
Input / 1M $0.050
Blend / 1M $0.088
Estimated monthly cost $87.00
OpenAI

GPT-5 Nano

Run it
Input / 1M $0.050
Blend / 1M $0.138
Estimated monthly cost $144.60
Gemini

Gemini 2.5 Flash-Lite

Run it
Input / 1M $0.100
Blend / 1M $0.175
Estimated monthly cost $169.20

Provider coverage

Compact provider coverage makes it easier to scan who is live, how much depth exists, and where pricing gaps still remain. Batch counts here mean listed batch pricing fields, using the same contract as the providers page.

Provider Status Models Lowest input Lowest output Cached Batch listed Updated
Anthropic ANTHROPIC
Live snapshot 4 $0.800 $4.00 4
0
Models with normalized batch pricing
Updated Mar 30, 02:54
No recent delta yet
Gemini GEMINI
Live snapshot 3 $0.100 $0.400 3
3
Models with normalized batch pricing
Updated Mar 30, 02:54
No recent delta yet
OpenAI OPENAI
Live snapshot 9 $0.050 $0.200 9
0
Models with normalized batch pricing
Updated Apr 12, 22:05
Latest change Apr 12, 22:05

Batch coverage on home and providers pages counts models with a normalized batch pricing field currently listed. It does not claim full provider-wide batch availability.

Latest price changes

40 tracked change events so far. Latest event: Apr 12, 22:05.

Open changes feed
OpenAI

GPT-5

-50%

input_price_per_million changed from USD 1.25 / 1M to USD 0.625 / 1M

$1.25 -> $0.625
Occurred Apr 10, 22:05
gpt-5
Inspect full history
OpenAI

GPT-5

-50%

output_price_per_million changed from USD 10 / 1M to USD 5 / 1M

$10.00 -> $5.00
Occurred Apr 10, 22:05
gpt-5
Inspect full history