14 Cheapest Llama Models

Every active Llama model ranked by blended price per million tokens. · verified 2026-10-23

#1 is Llama 3.2 1B Instruct (instruct) at $0.02 input and $0.02 output per million tokens.

// ranking

#modellabin $/1Mout $/1Mctx
1Llama 3.2 1B Instruct (instruct)Meta$0.02$0.02131K
2Llama-3.1-8B-InstructMeta$0.02$0.0516.4K
3Llama 3.2 1B InstructMeta$0.027$0.20160K
4Llama 3.2 3B InstructMeta$0.0509$0.33580K
5Muse Spark 1.2 ContributorMeta$0.10$0.201.05M
6Muse Spark 1.3 ContributorMeta$0.10$0.201.05M
7Llama 4 Scout 17BMeta$0.10$0.30327.7K
8Llama 3.3 70B TurboMeta$0.10$0.32131.1K
9Llama Guard 4 12BMeta$0.18$0.18163.8K
10Llama 3.2 11B Vision InstructMeta$0.0485$0.676128K
11Llama Guard 3 8BMeta$0.484$0.03131.1K
12Muse Spark 1.1Meta$1.25$4.251.05M
13Muse Spark OpenMeta$1.25$4.251.05M
14Muse WonderMeta$4.00$20.001.05M

// how this ranking works

Every active Llama model ranked by blended price per million tokens. Only active models with a published price and context window are included. Prices are base-tier list prices in USD per million tokens; the blended price weights input and output 3:1, a typical mix for chat and RAG workloads. The data comes from models.dev and is refreshed every day, so this ranking changes as soon as a provider changes its prices.

// other rankings

Cheapest LLM APIsCheapest Vision ModelsCheapest Reasoning ModelsAI Models With the Longest Context WindowCheapest Models With a 1M-Token ContextCheapest Open-Weights ModelsCheapest Models With Tool CallingCheapest Models With Structured OutputCheapest GPT AlternativesCheapest Claude AlternativesCheapest Gemini AlternativesNewest AI ModelsCheapest GPT (OpenAI) ModelsCheapest Claude ModelsCheapest Gemini & Gemma ModelsCheapest Grok ModelsCheapest Mistral ModelsCheapest DeepSeek ModelsCheapest Qwen ModelsCheapest GLM ModelsCheapest Kimi ModelsCheapest MiniMax ModelsCheapest Cohere Models