AIHubMix

A unified OpenAI-compatible API gateway with many models; 27+ models ending in -free are platform-subsidized and free within per-model caps (usually no credit card). One Key to compare GPT / Gemini / GLM / Kimi and try free image gen. Official post: Free AI models on AIHubMix.

Quotas, model lists, and “new-user free tries” change — trust aihubmix.com/models and the console.

Sites and entry points

Resource	URL
Site / sign-up	aihubmix.com
Free models post	docs.aihubmix.com/cn/blogs/free-ai-models
Model plaza	aihubmix.com/models
Developer docs	docs.aihubmix.com/cn
API Base URL	https://aihubmix.com/v1

How to get free access

1. Open aihubmix.com and register with email or OAuth (officially no credit card).
2. API Keys → create a Key (often sk-…).
3. In the model plaza filter free models and note IDs (usually end with -free).
4. In MiniTavern use Custom (OpenAI) / Custom (OpenAI Images) with Base URL + Key.

Item	Details
Free path	Model IDs with -free; RPM / daily token limits; quotas often reset daily (see each model page)
New users	Post claims limited free tries on some models after signup (e.g. GPT-5.5, GPT-Image-2); topping up unlocks more — follow the console
Card required	Free tier usually no
vs OpenRouter	OpenRouter free layer favors open-source breadth; AIHubMix stresses subsidized frontier closed-source *-free (e.g. GPT / Gemini free mirrors)

Model list

Excerpt from the official free-models post (~2026-05); live list follows the plaza. IDs are case-sensitive.

Text (chat / reasoning)

Chat Completions. MiniTavern: Custom (OpenAI).

General chat

Model ID	Notes
gpt-5.5-free	Flagship chat / Agent / code (platform-subsidized)
gpt-4o-free	Multimodal vision
gpt-4.1-free	Long context (~1M)
gpt-4.1-mini-free / gpt-4.1-nano-free	Speed / cost oriented
gemini-3-flash-preview-free	Very long context, multimodal
glm-4.7-flash-free	Zhipu Flash
mimo-v2-flash-free / ling-2.6-flash-free	Low-latency chat

Coding-oriented

Model ID	Notes
coding-glm-5.1-free	Officially highlighted free coding tier
coding-glm-5-free / coding-glm-5-turbo-free / coding-glm-4.7-free, etc.	GLM coding series
kimi-for-coding-free / k2.6-code-preview-free	Kimi coding
coding-minimax-m2-free … coding-minimax-m2.7-free	MiniMax coding series
step-3.5-flash-free	Lightweight completion

Agent / reasoning

Model ID	Notes
xiaomi-mimo-v2-pro-free	Tool calling / Agent
xiaomi-mimo-v2.5-free	Longer context and multi-tool calling

For roleplay chat, try gpt-4o-free, gemini-3-flash-preview-free, or glm-4.7-flash-free first; coding models lean toward code completion and may not suit long stories.

Images (text-to-image / edit)

Model ID	API note	App
gpt-image-2-free	POST /v1/images/generations (also edits)	Custom (OpenAI Images)
gemini-3.1-flash-image-preview-free (Nano Banana 2)	Officially may use Chat Completions for chat image gen	Image link may not be compatible; prefer Images endpoint or gpt-image-2-free

Video

That free list focuses on LLM + image; if video exists, follow plaza labels. MiniTavern has no built-in video generation entry.

How to fill in MiniTavern

General steps: Model link.

Chat

Path: Settings → Models → Model link → New

1. Platform: Custom (OpenAI)
2. Base URL: https://aihubmix.com/v1
3. Paste API Key
4. Model ID: gpt-4o-free or another plaza *-free
5. Test → on success, set as default

Image gen

Path: Settings → Image gen → Image gen link

1. Platform: Custom (OpenAI Images)
2. Same Base URL
3. Same Key; model gpt-image-2-free
4. Test, then use /imagine or image-gen entries

Related

• Free cloud overview
• Free image gen
• OpenCode Zen — coding-oriented curated gateway and another Free text set
• Agnes AI
• OpenRouter — open-source :free pool
• Model link