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 Nov 4, 2026 · 34 of 36 models publish this figure.

  1. 01
    $0.20USD / M tokens
  2. 02
    $0.20USD / M tokens
  3. 03
    $0.20USD / M tokens
#ModelUSD / M tokensCost-benefitContext
01Claude Haiku 5.5Anthropic$0.20247.0 pts/USD1.00M
02GPT-6 LunaOpenAI$0.20215.5 pts/USD1.05M
03Ministral 3 · 14BMistral AI$0.20—256K
04Qwen3.8 FlashAlibaba Cloud$0.23207.8 pts/USD—
05GLM-5.3 FlashZ.ai (Zhipu AI)$0.2375203.4 pts/USD—
06Mistral Small 4Mistral AI$0.2625—256K
07MiniMax M3.1 FlashMiniMax$0.2625184.4 pts/USD—
08Codestral · 25.08Mistral AI$0.45—128K
09GPT-5.6 LunaOpenAI$0.4597.8 pts/USD1.05M
10DeepSeek V4.1 FlashDeepSeek$0.52597.0 pts/USD1.00M
11MiniMax M3MiniMax$0.525——
12Gemini 3.5 Flash-LiteGoogle$0.85—1.05M
13Mistral Large 4Mistral AI$1.0544.7 pts/USD1.00M
14Gemini 3.8 FlashGoogle$1.5032.6 pts/USD1.05M
15DeepSeek V4 ProDeepSeek$1.9825.0 pts/USD1.00M
16GPT-6.1 TerraOpenAI$2.0023.8 pts/USD1.05M
17Claude Haiku 4.5Anthropic$2.00—200K
18Muse Spark 1.3Meta$2.0026.0 pts/USD—
19GLM-5.3Z.ai (Zhipu AI)$2.1523.5 pts/USD—
20Qwen3.8 MaxAlibaba Cloud$3.0016.8 pts/USD—
21Grok 4.7xAI$3.0015.4 pts/USD500K
22Grok 4.6xAI$3.0014.8 pts/USD500K
23GPT-6.1 SolOpenAI$4.0013.3 pts/USD1.05M
24Gemini 4 ArgonGoogle$4.0013.6 pts/USD1.05M
25Claude Sonnet 5.5Anthropic$4.0014.5 pts/USD1.00M
26Claude Sonnet 5Anthropic$4.0011.7 pts/USD1.00M
27Kimi K3.1 ProMoonshot AI$4.0013.9 pts/USD1.05M
28GPT-5.6 TerraOpenAI$4.5010.9 pts/USD1.05M
29Kimi K3Moonshot AI$6.008.6 pts/USD1.05M
30Claude Opus 5.5Anthropic$8.007.5 pts/USD1.00M
31GPT-5.6 SolOpenAI$8.006.6 pts/USD1.05M
32Claude Opus 5Anthropic$10.005.3 pts/USD1.00M
33GPT-6 AstraOpenAI$20.002.9 pts/USD1.05M
34Claude Fable 5.1Anthropic$20.002.6 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 Claude Haiku 5.5 vs GPT-6 LunaEstimate my monthly costGet a recommendation

Frequently asked questions

Which is the best AI model for price?

Claude Haiku 5.5 ($0.20) leads this ranking, followed by GPT-6 Luna ($0.20) · Ministral 3 · 14B ($0.20). Data reviewed on Nov 4, 2026.

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

Claude Haiku 5.5: 247.0 LLM Stats index points per dollar, at a blended price of $0.20 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.