Yandex Open-Sources AliceAI 80B MoE Model with 256K Context

Yandex open-sourced AliceAI-Foundation-80B-A3B-Base on Hugging Face: an 80B-parameter MoE model with 3B active per token and 256K context, positioned against Qwen and DeepSeek.

2026-09-21 ~ 2026-09-22 · 3 related posts

• Yandex open-sources 80B-A3B MoE base model AliceAI-Foundation, trending on Hugging Face — yandex · 2026-09-21
• Yandex open-sources 80B MoE Alice model: 3B active params, 256K context — teortaxesTex · 2026-09-22
• Yandex releases open AliceAI-Foundation-80B-A3B-Base, aimed at Qwen and DeepSeek Flash — Iwaku_Real · 2026-09-22