Aion 3.5
aion-3-5aion-3.5aion-labs-aion-3-5- 🎯 Multi-model roleplay and storytelling system from Aion Labs
- 📏 262K token context for long-running narratives
- 🧩 Several specialized models collaborate on each response
- 🔧 Function calling, reasoning, and web search supported
- 🎨 Handles mature and darker themes with nuance
Aion Labs is an Israeli venture studio working at the intersection of artificial intelligence and pharmaceutical research. The group focuses on applying machine learning to drug discovery and development, partnering with scientific and industry collaborators to translate AI…
Explore 4 more models by AION Labs →Aion 3.5 takes an unusual approach to interactive fiction: rather than a single checkpoint, it orchestrates multiple specialized models that collaborate on every response, each contributing to narrative structure, pacing, tension, and conflict. Built by Aion Labs on top of the GLM family of models, it arrived in September 2026 as the full-size entry in the company's roleplay-oriented lineup.
Within that lineup, Aion 3.5 is the high-capability tier, sitting above the lighter, faster Aion 3.5 Mini and alongside the earlier Aion 3.0 and Aion 3.0 Mini pairing. Its 262,144-token context window is generous enough to hold extended story arcs, character sheets, and world-building notes in a single session, and reasoning, function calling, and web search are all available for richer, more dynamic scenes.
It is tuned to treat mature and darker themes with nuance rather than deflection, which makes it a fit for long-form creative writing, character-driven roleplay, game masters running persistent campaigns, and tool-augmented interactive fiction where the story needs to remember what happened fifty scenes ago.
This About section is AI-generated from public sources via VeniceStats + Venice inference, with no human editing. It may contain inaccuracies.
Daily usage · Aion 3.5 · Organic sellers
| Seller | Health | Rep↓ | Sales | Tokens | Routing | Input $/M | Cached $/M | Output $/M | Save |
|---|---|---|---|---|---|---|---|---|---|
| ▲ Apex Ant 0x73b4…e736 | — | 81.71 | $0.07 | 69.65k | #1 | $1.875 | $0.47 | $3.75 | −38% |
| attacompute.ai 0xd5e7…7be0 | healthy | 51.22 | $0.03 | 31.98k | gated | $3.31 | $0.83 | $6.61 | — |
| surplus-hive 0x6145…5c58 | — | 0.00 | — | — | gated | $9.1146 | $9.1146 | $18.2292 | — |
"Best price" and the seller table are live Antseed catalog data (advertised $/1M tokens — or $ per generated image for unit-billed image models — not settled amounts). Reputation = buyer trust score (0-100, the Antseed SDK's own formula). "Routing" = the SDK's default buyer routing (what the VPR desktop app ships with): a trust ≥ 60 gate on the effective reputation, then cheapest-first among routable sellers; live failover state (per-peer cooldowns) is buyer-side runtime and not included. Model knowledge (TLDR, provider, About) via the VeniceStats enrichment layer. Advertised catalog, not the model used in any specific purchase. "Usage on Antseed" counts only settlements whose buyers share the per-model split on-chain (metadata v2/v3, opt-in), so every usage figure is a lower bound.