INTELLECT-3

Open weights Active Text LLM MoE

106B MoE trained with SFT and large-scale RL on top of the GLM-4.5-Air base model.

Released
Nov 26, 2025
Context window
Not disclosed
Max output
Not disclosed
Parameters
106B 12B active
Input price
Not disclosed
Output price
Not disclosed

Prices are first-party list prices in USD at the standard tier. Last verified Oct 30, 2026.

Pricing

No list price recorded for this model yet.

Specifications

Developer
Prime Intellect
Model type
Text LLM
Architecture
MoE
Parameters
106B (12B active)
Input
Text
Output
Text
Knowledge cutoff
Not disclosed
License
MIT License
API access
First-party API
Capabilities
Reasoning