KT: Mi:dm K 3.0 Pro
kt · Text
Mi:dm K 3.0 Pro is a Mixture-of-Experts reasoning model from KT, with 120B total and 12B active parameters. It uses a hybrid architecture that mixes Mamba, MoE and attention layers.
Model ID kt/midm-k-3.0-proUpdated Nov 5, 2026
Pricing & billing
Billing unit: 1M tokens. Vendor reference price; conversion is approximate.
ItemUSD referenceList price (USD)
Input$0.20$0.20
Output$0.80$0.80
Cache read—No public pricing—No public pricing
Cache write—No public pricing—No public pricing
Specifications
- Form
- Text model
- Input modalities
- Text
- Output modalities
- Text
- Context length
- 262K
- Max output
- 236K
- Released
- Nov 5, 2026
Uses and capabilities
Capabilities Text model、Function calling、Structured output、Reasoning
Function calling
Structured output
Reasoning
Prompt caching
Open weights
Sources
Use these links to verify specs and pricing — they may point to the vendor's site or to a third-party catalog.