Live survey286 models · 13 providers · 19,992 auditsAudited verification-before-completion · SAFE →What changedHow we verifyllms.txt

286 models · 13 providers · prices checked 12 Oct 2026

AI model API pricing

What 180 models cost per 1M input and output tokens, by provider and by tier — standard, batch, free and more. Each price opens its source and the date we retrieved it.

Median input / 1M
$0.36

Paid models, standard tier

Median output / 1M
$1.92

Paid models, standard tier

Cheapest paid
$0.02
derived We computed
Source
Our catalogue
Retrieved
10 Oct 2026
Note
Lowest standard-tier input price across our catalogue.

gpt-oss-20b · input

Free models
10

See free models

Ask an AI about this pageGChatGPTCClaudePPerplexityGeGeminiXGrok
More
CoCopilotAIGoogle AI Mode
Cite

Cite this page

ModelsAtlas. “AI model API pricing”. Retrieved 12 Oct 2026. https://modelsatlas.com/pricing

Pricing by provider

All 13 providers →
Qwen (Alibaba) 35From $0.03 · Qwen3.7 FlashOpenAI 35From $0.02 · gpt-oss-20bGoogle 25From $0.05 · Gemma 3 4BMeta 14From $0.03 · Llama 3.2 1B InstructMistral AI 18From $0.03 · Mistral NemoAnthropic 17From $0.10 · Claude Haiku 5.5NVIDIA 5From $0.06 · Nemotron 3 Nano 30B A3BDeepSeek 9From $0.08 · DeepSeek V4 Flash 0731Microsoft 2From $0.07 · Phi 4MiniMax 8From $0.20 · MiniMax-01xAI 7From $1.00 · Grok Build 0.1Cohere 5From $0.04 · Command R7B (12-2024)

Pricing by tier

The same model, billed differently

Standard 180 offersThe headline pay-as-you-go rateBatch 52 offersAsynchronous requests, billed below the standard rateFree 8 offersListed at no per-token chargeThinking 3 offersExtended-reasoning variants of a model

Cheapest model for each capability

Cheapest models →
You needCheapest paid modelInput / 1MOutput / 1M
1M contextQwen3.7 FlashQwen (Alibaba)$0.03$0.13
Open weightsQwen2.5 7B InstructQwen (Alibaba) · open weights$0.10$0.20
Extended thinkingQwen3 235B A22B Instruct 2507Qwen (Alibaba)$0.09$0.55
Standard tier. A model counts only for what its record lists (input types, context, weights, tiers); one with nothing recorded is not counted.

The 50 lowest-priced paid models

Standard tier · input price, low to high

#ModelInput / 1MOutput / 1MContext
1gpt-oss-20bOpenAI$0.02$0.09131KAlternatives
2Llama 3.2 1B InstructMeta$0.03$0.2060KAlternatives
3Mistral NemoMistral AI$0.03$0.03131KAlternatives
4Qwen3.7 FlashQwen (Alibaba)$0.03$0.131MAlternatives
5gpt-oss-120bOpenAI$0.03$0.15131KAlternatives
6Command R7B (12-2024)Cohere$0.04$0.15128KAlternatives
7Llama 3.1 8B InstructMeta$0.05$0.08131KAlternatives
8Mistral Small 3Mistral AI$0.05$0.0833KAlternatives
9Gemma 3 4BGoogle$0.05$0.10131KAlternatives
10Gemma 3 12BGoogle$0.05$0.15131KAlternatives
11Llama 3.2 3B InstructMeta$0.05$0.33131KAlternatives
12Nemotron 3 Nano 30B A3BNVIDIA$0.06$0.24262KAlternatives
13Qwen3.5-FlashQwen (Alibaba)$0.07$0.261MAlternatives
14Gemma 4 26B A4BGoogle$0.07$0.23262KAlternatives
15Phi 4Microsoft$0.07$0.1416KAlternatives
16Nemotron 3.5 Lightning (free)NVIDIA$0.07$0.201MAlternatives
17Qwen3 Coder 30B A3B InstructQwen (Alibaba)$0.07$0.28262KAlternatives
18gpt-oss-safeguard-20bOpenAI$0.08$0.30131KAlternatives
19Qwen3 32BQwen (Alibaba)$0.08$0.28131KAlternatives
20Gemma 3 27BGoogle$0.08$0.45131KAlternatives
21Nemotron 3 SuperNVIDIA$0.08$0.45262KAlternatives
22Qwen3.5-35B-A3BQwen (Alibaba)$0.08$0.75262KAlternatives
23DeepSeek V4 Flash 0731DeepSeek$0.08$0.171.05MAlternatives
24Gemma 4 31BGoogle$0.09$0.34262KAlternatives
25Qwen3 235B A22B Instruct 2507Qwen (Alibaba)$0.09$0.55262KAlternatives
26Mistral Small 3.2 24BMistral AI$0.09$0.25256KAlternatives
27Ministral 3 3B 2512Mistral AI$0.10$0.10131KAlternatives
28Qwen3.5-9BQwen (Alibaba)$0.10$0.15262KAlternatives
29Muse Spark 1.2 ContributorMeta$0.10$0.201.05MAlternatives
30Muse Spark 1.3 ContributorMeta$0.10$0.201.05MAlternatives
31Qwen2.5 7B InstructQwen (Alibaba) · open weights$0.10$0.2033KAlternatives
32Llama 4 ScoutMeta$0.10$0.301.31MAlternatives
33Qwen3 30B A3B Instruct 2507Qwen (Alibaba)$0.10$0.30262KAlternatives
34Voxtral Small 24B 2507Mistral AI$0.10$0.3033KAlternatives
35Claude Haiku 5.5Anthropic$0.10$0.501MAlternatives
36GPT-6 LunaOpenAI$0.10$0.501.05MAlternatives
37Qwen3.6 35B A3BQwen (Alibaba) · open weights$0.10$0.95262KAlternatives
38Qwen3 Next 80B A3B InstructQwen (Alibaba)$0.10$1.10262KAlternatives
39Qwen3 14BQwen (Alibaba) · open weights$0.12$0.2441KAlternatives
40Qwen3 30B A3BQwen (Alibaba)$0.12$0.5041KAlternatives
41Qwen3 Coder NextQwen (Alibaba)$0.12$0.80262KAlternatives
42Ministral 3 8B 2512Mistral AI$0.15$0.15262KAlternatives
43Qwen3.8 FlashQwen (Alibaba)$0.15$0.471MAlternatives
44Qwen3.8 Omni FlashQwen (Alibaba)$0.15$0.471MAlternatives
45Command R (08-2024)Cohere$0.15$0.60128KAlternatives
46Mistral Small 4Mistral AI$0.15$0.60262KAlternatives
47Qwen3 VL 30B A3B InstructQwen (Alibaba)$0.15$0.60262KAlternatives
48Llama Guard 4 12BMeta$0.18$0.18164KAlternatives
49Llama 4 MaverickMeta$0.19$0.651.05MAlternatives
50Qwen3.6 FlashQwen (Alibaba)$0.19$1.131MAlternatives
Cheapest paid models first. Free models are on the free tier page.
Showing 1–50 of 178 paid models · all 180 priced models

Pricing questions

How is AI API pricing calculated?
Per token, with separate rates for input (your prompt) and output (the reply), quoted per 1 million tokens. 1M tokens is roughly 750,000 English words.
What is batch pricing?
Batch requests are processed asynchronously and billed below the standard rate. 52 models list a batch offer in our catalogue; where the same model lists both, batch is a median 50% of its standard price.
How current are these prices?
The newest price on this page was retrieved 12 Oct 2026. Each price shows the date we retrieved it; open a price marked src to see its source.
What is the cheapest paid AI model?
gpt-oss-20b (OpenAI) at $0.02 per 1M input tokens and $0.09 per 1M output on the standard tier. 10 models are listed at no per-token charge.

Keep going in your AI

Opens with this page as the source. We add nothing else to the prompt.