Catalog data · Updated July 24, 2026

Cheapest LLM APIs by Token Price

Compare active text-generation API offers by their known input and output token prices. Free, promotional, and missing prices are kept out of this paid-price list.

Objective catalog fields No overall quality score
856
Paid text API offers
24
Providers shown
1
Data sections
6,154
Catalog records

Lowest input-token prices

Paid text-generation offers with known input and output token prices, ordered by input price.

Showing 6 of 856

Model Provider offers Why listed Context Input / output price Updated
OpenAI: GPT-5.5 Pro
openai/gpt-5.5-pro
Openrouter
1 offer
Known paid token prices 1,050,000
$30.00 / $180.00
per 1M tokens
April 23, 2026
openai/gpt-5.5-pro
openai/gpt-5.5-pro
Zenmux
1 offer
Known paid token prices 1,050,000
$30.00 / $180.00
per 1M tokens
April 23, 2026
GPT-5.4 Pro
openai-gpt-54-pro
Venice
1 offer
Known paid token prices 1,000,000
$37.50 / $225.00
per 1M tokens
June 11, 2026
GPT-5.5 Pro
openai-gpt-55-pro
Venice
1 offer
Known paid token prices 1,000,000
$37.50 / $225.00
per 1M tokens
June 11, 2026
openai/gpt-5.4-pro
openai/gpt-5.4-pro
Zenmux
1 offer
Known paid token prices 1,050,000
$45.00 / $225.00
per 1M tokens
March 20, 2026
OpenAI: o1-pro
openai/o1-pro
Openrouter
1 offer
Known paid token prices 200,000
$150.00 / $600.00
per 1M tokens
March 19, 2025

How the price list works

The page compares provider offers with active text execution and positive input and output token prices.

Inclusion criteria

  • The offer passes the active text-generation rules used by the LLM models directory.
  • The offer has positive numeric values for both input-token and output-token prices.

Exclusions

  • The list excludes zero prices because they can be free tiers, promotions, or incomplete records.
  • The list excludes offers with a missing input or output token price.

Data rules

  • Comparison unit: The table shows catalog token prices per one million tokens.
  • Ordering: Offers are ordered by input price, then output price, provider ID, and model ID.
  • Offer scope: Each row is a provider offer. The page does not merge prices from different providers.

Limits of the data

  • The lowest input price is not always the lowest total price for a specific workload.
  • Provider discounts, cache prices, batches, subscriptions, and minimum charges can change effective cost.
  • A low price does not measure model quality, speed, or fitness for a task.

Use this list

Start with input price when your workload sends much more text than it receives. Check output price when responses are long. Follow each model link to inspect its complete provider record before you select an API.

Sources

  • llm_db — Source for active provider offers, typed execution data, and token prices. Checked 2026-07-30.