InterconnectsInterconnects Artifacts Hub
withVAILVAIL
All models

Instella-MoE-16B-A3B-Think

byAMDAMD· 24 Jul 2026
General purpose

A 16B-A3B MoE trained by AMD on Instinct cards. AMD also provides all the different stages, from the base to the SFT checkpoints, as well as MidTrain and DPO.

Specs
Params16B (2.8B active)
LicenseResearchRAIL
Tags
models/llms
Resources
Hugging Face — amd/Instella-MoE-16B-A3B-ThinkPaper / tech reportAppeared in Artifacts Log #23
How these metrics work
Adoption · Hugging Face
RAM score
—
Hugging Face Downloads
2K
last 30d
2K
all time
HF Likes
145

Relative Adoption Metric contextualizes downloads against the model's size bucket.

Related Models

More from AMD

Instella-3B
AMDAMD05 Mar 2025
AMD-OLMo-1B
AMDAMD31 Oct 2024

Similar Models

Gravity-16B-A3B-Base
Trillion LabsTrillion Labs09 Apr 2026
16B params
Nemotron-Labs-Diffusion-14B
NVIDIANVIDIA22 Apr 2026
14B params
Mellum2-12B-A2.5B-Instruct
JetBrainsJetBrains26 May 2026
12B params
Ling-mini-2.0
Inclusion AIInclusion AI08 Sep 2025
16B params