Aion 3.5 is a multi-model roleplaying and storytelling system from AionLabs, built on the GLM family of models.
It uses a collaborative generation process in which multiple specialized models each contribute to a response, producing stronger narrative structure and more compelling tension and conflict.
| Developer | AionLabs |
|---|---|
| Model ID | aion-labs/aion-3.5 |
| Release Date | September 23, 2026 |
| Context Window | 262,144 tokens (~262K) |
| Max Output Tokens | 32,768 tokens |
| Modalities | Text → Text |
| Capabilities | file input, parallel tool calls, reasoning, streaming, structured (JSON schema) output, tool/function calling, web search |
Model type
Multi-model roleplaying and storytelling system.
Developer
AionLabs.
Architecture
Built on the GLM family of models.
Generation approach
Uses collaborative generation, where multiple specialized models contribute to one response.
Focus
Designed to produce stronger narrative structure and more compelling tension and conflict.
Roleplaying
Suited to interactive roleplaying experiences.
Storytelling
Suited to narrative generation and story development.
Narrative writing
Useful for outputs that benefit from stronger structure and dramatic tension.
Aion 3.5 Mini
AionLabs also offers Aion 3.5 Mini, which is described as the smaller, lower-cost sibling of Aion 3.5 and uses the same collaborative generation process.
Sources: https://openrouter.ai/aion-labs/aion-3.5 · https://openrouter.ai/provider/aion-labs · https://openrouter.ai/aion-labs
Measured on the AI/ML API gateway across live traffic. Figures are medians unless stated. Last updated 2026-10-12.
Aion-3.5 is documented as suited for roleplaying and storytelling, where its collaborative generation process is used to produce stronger narrative structure and more compelling tension and conflict.
It is also documented for richer interactive fiction, with support for tool calling noted in that context.
It is further described as handling mature and darker themes with nuance.
Another documented use case is chat-style use through a frontend for power users that emphasizes character creation, roleplay, and prompt customization.
import requests r = requests.post( "https://api.aimlapi.com/v1/chat/completions", headers={"Authorization": "Bearer " + AIMLAPI_KEY}, json={ "model": "aion-labs/aion-3.5", "messages": [ { "role": "user", "content": "Hello!" } ] }, ) print(r.json())
const r = await fetch("https://api.aimlapi.com/v1/chat/completions", { method: "POST", headers: { Authorization: `Bearer ${process.env.AIMLAPI_KEY}`, "Content-Type": "application/json", }, body: JSON.stringify({ "model": "aion-labs/aion-3.5", "messages": [ { "role": "user", "content": "Hello!" } ] }), }); console.log(await r.json());
curl -X POST https://api.aimlapi.com/v1/chat/completions \ -H "Authorization: Bearer $AIMLAPI_KEY" \ -H "Content-Type: application/json" \ -d '{"model":"aion-labs/aion-3.5","messages":[{"role":"user","content":"Hello!"}]}'
OpenAI-compatible — swap the base URL and it works with your existing SDK.
| Type | Price |
|---|---|
| Input | |
| Output | |
| Cached input |
A documented caveat is that Aion Labs offers a free tier with request and token limits, including 15 requests per minute and 20,000 tokens per minute/day on the free tier.
Aion-3.5 is a multi-model roleplaying and storytelling system from AionLabs built on the GLM family of models.
It uses a collaborative generation process where multiple specialized models each contribute to a response. This is intended to produce stronger narrative structure and more compelling tension and conflict.
It is optimized for roleplaying and storytelling tasks. The model is focused on narrative quality rather than general-purpose chat alone.
Aion-3.5 is a multi-model system. Multiple specialized models work together during generation.
Aion-3.5 is built on the GLM family of models.
Yes, the model supports tools. It can work with function-style tool use.
Yes, structured output is supported for this model.
Yes, streaming is supported.
Yes, parallel tool calls are supported.
Yes, file input and web search are supported.
