Subscribe
Ranking · AI models

Cheapest AI models per million tokens

Standard API price blended 3:1 (input:output), without caching or batch discounts.

Data reviewed on Sep 8, 2026 · 17 of 17 models publish this figure.

  1. 01
    $0.20USD / M tokens
  2. 02
    Mistral Small 4Mistral AI
    $0.2625USD / M tokens
  3. 03
    $0.45USD / M tokens
#ModelUSD / M tokensCost-benefitContext
01Ministral 3 · 14BMistral AI$0.20—256K
02Mistral Small 4Mistral AI$0.2625—256K
03Codestral · 25.08Mistral AI$0.45—128K
04GPT-5.6 LunaOpenAI$0.45—1.05M
05DeepSeek V4 FlashDeepSeek$0.66—1.00M
06Mistral Large 3Mistral AI$0.75—256K
07Gemini 3.8 FlashGoogle$1.5034.0 pts/USD1.05M
08DeepSeek V4 ProDeepSeek$1.98—1.00M
09Claude Haiku 4.5Anthropic$2.00—200K
10Mistral Medium 3.5Mistral AI$3.00—256K
11Grok 4.6xAI$3.00—500K
12Claude Sonnet 5Anthropic$4.00—1.00M
13GPT-5.6 TerraOpenAI$4.5011.4 pts/USD1.05M
14GPT-5.6 SolOpenAI$8.006.9 pts/USD1.05M
15Claude Opus 5Anthropic$10.005.5 pts/USD1.00M
16GPT-6 AstraOpenAI$20.003.0 pts/USD1.05M
17Claude Fable 5.1Anthropic$20.002.8 pts/USD1.00M

How it is calculated

  • The indexes (overall, reasoning, coding) are the ones LLM Stats publishes; Codifly does not run these tests.
  • Prices are each provider's official standard API prices, without caching or batching.
  • The blended price weighs 3 input tokens per output token; cost-benefit divides the overall index by that price.
Methodology and sources

Which one fits your case?

The top of a ranking is not always the most cost-effective at your volume. Estimate your cost or ask us for a written recommendation.

Compare Ministral 3 · 14B vs Mistral Small 4Estimate my monthly costGet a recommendation

Frequently asked questions

Which is the best AI model for price?

Ministral 3 · 14B ($0.20) leads this ranking, followed by Mistral Small 4 ($0.2625) · Codestral · 25.08 ($0.45). Data reviewed on Sep 8, 2026.

Which model in this ranking has the best price-performance?

Gemini 3.8 Flash: 34.0 LLM Stats index points per dollar, at a blended price of $1.50 per million tokens.

How is this ranking calculated?

The indexes (overall, reasoning, coding) are the ones LLM Stats publishes; Codifly does not run these tests. Prices are each provider's official standard API prices, without caching or batching. The blended price weighs 3 input tokens per output token; cost-benefit divides the overall index by that price.

The Codifly brief

A clearer perspective.
In your inbox.

Analysis, guides and new technology comparisons.

You can unsubscribe whenever you like.