Skip to content
Sovyron

Cheapest LLM APIs

As of Nov 22, 2026, the cheapest LLM API is GLM-5.3-Flash at $0.0062 per 1M input and $0.0185 per 1M output tokens on JoinGonka.

The table ranks the 20 cheapest of 429 qualifying models offered by at least 3 providers, by the cheapest offer's 3:1 input:output blended price. Every model Sovyron compares.

Cheapest LLM APIs, USD per 1M tokens, as of Nov 22, 2026
#Model
1GLM-5.3-Flash$0.0062JoinGonka$0.0185JoinGonka$0.00931.0M113
2DeepSeek V4 Flash 0731$0.0062JoinGonka$0.0185JoinGonka$0.00931.1M53
3MiniMax-M2.7$0.0062JoinGonka$0.0185JoinGonka$0.0093205k32
4Llama 3.2 1B Instruct$0.01Inference$0.01Inference$0.01131k3
5Llama Prompt Guard 2 22M$0.01Helicone$0.01Helicone$0.015123
6Llama 3.2 3B Instruct$0.02Eden AI$0.02Eden AI$0.02131k6
7Ling 3.0 Tiny$0.015Kilo Gateway$0.045Kilo Gateway$0.0225262k3
8Llama-3.1-8B-Instruct$0.02Kilo Gateway$0.04Kilo Gateway$0.025131k10
9DeepSeek V4.1 Flash 1017$0.01Refuseless$0.08Refuseless$0.02751.1M73
10GPT-5.6 Luna$0.0123xxapi$0.0738xxapi$0.02771.1M54
11Mistral Nemo$0.029OpenRouter$0.03OpenRouter$0.0293131k8
12GPT OSS 20B$0.015SOKKAN Router$0.075SOKKAN Router$0.03131k38
13Ling 3.0 Flash$0.021OpenRouter$0.063OpenRouter$0.0315262k8
14MiniMax-M3.1-Flash$0.02Refuseless$0.08Refuseless$0.0351.0M34
15Nemotron 3.5 Lightning 30B A3B$0.02Kilo Gateway$0.096Kilo Gateway$0.0391.0M7
16Ministral 3B$0.04Azure$0.04Azure$0.04131k7
17Gemma 3 4B IT$0.04Kilo Gateway$0.05Kilo Gateway$0.0425131k6
18Qwen3.5 4B$0.04EmpirioLabs AI$0.07EmpirioLabs AI$0.0475262k3
19Qwen3.7 Flash$0.0282AIHubMix$0.1128AIHubMix$0.04931.0M14
20GLM Flash Latest$0.0399OpenRouter$0.0784OpenRouter$0.04951.0M3

FAQ

What is the cheapest LLM API?

As of Nov 22, 2026, the cheapest LLM API is GLM-5.3-Flash at $0.0062 per 1M input and $0.0185 per 1M output tokens on JoinGonka.

Which LLM APIs are cheapest after GLM-5.3-Flash?

DeepSeek V4 Flash 0731 ($0.0093 blended per 1M tokens on JoinGonka), MiniMax-M2.7 ($0.0093 blended per 1M tokens on JoinGonka), Llama 3.2 1B Instruct ($0.01 blended per 1M tokens on Inference), as of Nov 22, 2026.

How is this list ranked?

By each model's cheapest metered offer, blended 3:1: (3 × input + output) ÷ 4 per 1M tokens. Free tiers and subscription plans are excluded, and a model needs at least 3 paid providers to qualify.

Other cheapest lists