# Cheapest Chat API Models

As of October 21, 2026, among LiteLLM's active models with a 32K-token context window or more, offered by at least two providers or by their own lab, the cheapest of the chat models is Qwen2.5 Coder 3B Instruct, at $0.015 per 1M tokens (blended), from Nscale.

- Page: https://models.litellm.ai/rankings/cheapest-chat-models
- How it's sorted: Sorted by blended price: 3 input tokens for every output token, at each model's headline price (its own lab's API where it sells it, otherwise its cheapest provider). Only active models with a 32K-token context window or more, offered by at least two providers or by their own lab. Free models are left out.

| # | Model | Lab | Blended price per 1M tokens | Price per 1M (in / out) | Context | Providers |
|---|---|---|---|---|---|---|
| 1 | [Qwen2.5 Coder 3B Instruct](https://models.litellm.ai/models/qwen2.5-coder-3b-instruct) | Alibaba | $0.015 | $0.01 / $0.03 | 32K | 2 |
| 2 | [Qwen2.5 Coder 7B](https://models.litellm.ai/models/qwen2.5-coder-7b) | Alibaba | $0.015 | $0.01 / $0.03 | 32K | 3 |
| 3 | [Qwen2.5 Coder 7B Instruct](https://models.litellm.ai/models/qwen2.5-coder-7b-instruct) | Alibaba | $0.015 | $0.01 / $0.03 | 32K | 2 |
| 4 | [Llama3.2 11B Vision Instruct](https://models.litellm.ai/models/llama-3.2-11b-vision-instruct) | Meta | $0.0175 | $0.015 / $0.025 | 128K | 5 |
| 5 | [Llama3.2 3B Instruct](https://models.litellm.ai/models/llama-3.2-3b-instruct) | Meta | $0.0175 | $0.015 / $0.025 | 128K | 8 |
| 6 | [Llama 3.2 1B Instruct](https://models.litellm.ai/models/llama-3.2-1b-instruct) | Meta | $0.02 | $0.02 / $0.02 | 131K | 5 |
| 7 | [Mistral Nemo Instruct](https://models.litellm.ai/models/mistral-nemo-instruct) | Mistral AI | $0.02175 | $0.019 / $0.03 | 128K | 5 |
| 8 | [DeepSeek R1 Distill Llama 8B](https://models.litellm.ai/models/deepseek-r1-distill-llama-8b) | DeepSeek | $0.025 | $0.025 / $0.025 | 128K | 2 |
| 9 | [Llama 3.1 8B Instruct Turbo](https://models.litellm.ai/models/llama-3.1-8b-instruct-turbo) | Meta | $0.025 | $0.02 / $0.04 | 128K | 2 |
| 10 | [GPT OSS 20B](https://models.litellm.ai/models/gpt-oss-20b) | OpenAI | $0.02838 | $0.0145 / $0.07 | 128K | 16 |
| 11 | [Qwen2 VL 7B Instruct](https://models.litellm.ai/models/qwen2-vl-7b-instruct) | Alibaba | $0.03 | $0.02 / $0.06 | 128K | 2 |
| 12 | [Llama 3.1 8B](https://models.litellm.ai/models/llama-3.1-8b) | Meta | $0.035 | $0.03 / $0.05 | 128K | 5 |
| 13 | [Llama 3.2 3B](https://models.litellm.ai/models/llama-3.2-3b) | Meta | $0.05 | $0.04 / $0.08 | 128K | 3 |
| 14 | [Qwen2.5 7B Instruct](https://models.litellm.ai/models/qwen2.5-7b-instruct) | Alibaba | $0.055 | $0.04 / $0.10 | 32K | 3 |
| 15 | [Llama Guard 3 8B](https://models.litellm.ai/models/llama-guard-3-8b) | Meta | $0.055 | $0.055 / $0.055 | 128K | 5 |
| 16 | [Mistral Small 24B Instruct](https://models.litellm.ai/models/mistral-small-24b-instruct) | Mistral AI | $0.0575 | $0.05 / $0.08 | 32K | 3 |
| 17 | [Nova Micro](https://models.litellm.ai/models/nova-micro) | Amazon | $0.06125 | $0.035 / $0.14 | 128K | 2 |
| 18 | [Nova Micro V1](https://models.litellm.ai/models/nova-micro-v1) | Amazon | $0.06125 | $0.035 / $0.14 | 128K | 1 |
| 19 | [Gemma 3 12B IT](https://models.litellm.ai/models/gemma-3-12b-it) | Google | $0.0625 | $0.05 / $0.10 | 128K | 4 |
| 20 | [GPT OSS 120B](https://models.litellm.ai/models/gpt-oss-120b) | OpenAI | $0.065 | $0.03 / $0.17 | 131K | 24 |