Cheapest by list price
The cheapest LLM APIs right now
Every model below is sold by at least three providers; each row shows the cheapest one. Right now qwen2.5-coder-3b-instruct at Nscale is cheapest at $0.015 per 1M blended tokens.
Standard list prices in USD, ranked by blended price (3 input : 1 output tokens). Last price change Sep 28, 2026. How we rank.
Top 25 by price
Cheapest is not the same as best. Price says nothing about output quality, speed or rate limits, so check public benchmarks such as Artificial Analysis before you switch.
| # | Model | Cheapest provider | Input / 1M | Output / 1M | Blended / 1M | Context |
|---|---|---|---|---|---|---|
| 1 | qwen2.5-coder-3b-instruct | NscalePricing page | $0.01 | $0.03 | $0.015 | - |
| 2 | qwen2.5-coder-7b | Nebius Token FactoryPricing page | $0.01 | $0.03 | $0.015 | 33K |
| 3 | qwen2.5-coder-7b-instruct | NscalePricing page | $0.01 | $0.03 | $0.015 | - |
| 4 | Llama 3.2 11B Vision Instruct | LambdaPricing page | $0.015 | $0.025 | $0.0175 | 131K |
| 5 | Llama 3.2 3B Instruct | LambdaPricing page | $0.015 | $0.025 | $0.0175 | 131K |
| 6 | Llama 3.2 1B Instruct | Novita AIPricing page | $0.02 | $0.02 | $0.02 | 131K |
| 7 | mistral-nemo-instruct-2407 | DeepInfraPricing page | $0.019 | $0.03 | $0.0218 | 131K |
| 8 | deepseek-r1-distill-llama-8b | NscalePricing page | $0.025 | $0.025 | $0.025 | - |
| 9 | Llama-3.1-8B-Instruct | Novita AIPricing page | $0.02 | $0.05 | $0.0275 | 16K |
| 10 | Llama Guard 3 8B | Nebius Token FactoryPricing page | $0.02 | $0.06 | $0.03 | 128K |
| 11 | qwen2-vl-7b-instruct | Nebius Token FactoryPricing page | $0.02 | $0.06 | $0.03 | 131K |
| 12 | llama3-8b-instruct | DeepInfraPricing page | $0.03 | $0.06 | $0.0375 | 8K |
| 13 | Ministral 3B | Microsoft Azure AI FoundryPricing page | $0.04 | $0.04 | $0.04 | 128K |
| 14 | Gemma 3 4B IT | Amazon BedrockPricing page | $0.04 | $0.08 | $0.05 | 128K |
| 15 | Qwen2.5-7B-Instruct | SiliconFlowPricing page | $0.05 | $0.05 | $0.05 | 33K |
| 16 | gpt-oss-20b | W&B Inference (CoreWeave)Pricing page | $0.03 | $0.13 | $0.055 | 131K |
| 17 | Mistral Small 3 | DeepInfraPricing page | $0.05 | $0.08 | $0.0575 | 33K |
| 18 | Qwen3-8B | SiliconFlowPricing page | $0.06 | $0.06 | $0.06 | 131K |
| 19 | Gemma 3 12B IT | Novita AIPricing page | $0.05 | $0.10 | $0.0625 | 131K |
| 20 | Llama 4 Maverick 17B 128E Instruct FP8 | LambdaPricing page | $0.05 | $0.10 | $0.0625 | 131K |
| 21 | Llama 4 Scout 17B 16E Instruct | LambdaPricing page | $0.05 | $0.10 | $0.0625 | 16K |
| 22 | Qwen2.5 Coder 32B Instruct | LambdaPricing page | $0.05 | $0.10 | $0.0625 | 131K |
| 23 | qwen3-32b-fp8 | LambdaPricing page | $0.05 | $0.10 | $0.0625 | 131K |
| 24 | gpt-oss-120b | W&B Inference (CoreWeave)Pricing page | $0.03 | $0.17 | $0.065 | 131K |
| 25 | deepseek-r1-distill-qwen-14b | NscalePricing page | $0.07 | $0.07 | $0.07 | - |
Only models listed by at least three providers are ranked. Order depends on listed price only; no provider pays for its position.
Compare more
Pick a model to see every provider
The cheapest LLM APIs with a 1M-token context window - All LLM API prices - Recent price changes