Tool tag: MoE

Tencent Hy4 preview

Tencent's open-weight flagship, released August 28, 2026. A 770B MoE with 49B active, a 1M-token context, Apache 2.0 — and explicitly a preview.

GLM-5.3

Zhipu AI's current line: GLM-5.3, a post-trained coding and security flagship, and GLM-5.3-Flash, a 320B/18B multimodal MoE with hybrid attention.

Qwen3.8-Max

Alibaba's flagship Qwen model, released August 3, 2026. A 2.4T-parameter MoE with 95B active, 1M token context, text/image/video input — and open weights.

Kimi K3

Moonshot AI's flagship: a 2.8T-parameter MoE with 104B active, 1M context and native vision. Open weights shipped July 27, 2026.

Inkling

Thinking Machines Lab's first open-weights model: a 975B-parameter multimodal MoE with 41B active, 1M-token context and controllable thinking effort.

LongCat-2.0

Meituan's open-weight flagship, announced June 30, 2026. A 1.6T-parameter MoE with ~48B active parameters, a 1M-token context window, and an MIT license.

Nemotron 3 Ultra

NVIDIA's open Nemotron line: the 550B/55B-active Nemotron 3 Ultra flagship and Nemotron 3.5 Lightning, a 30B/3B-active agent-execution model, both OpenMDW.

MiniMax-M3

MiniMax's multimodal flagship, released June 1, 2026. A 428B-parameter MoE with ~23B active, a 1M-token context, and open weights under a community license.

Step-3.7-Flash

StepFun's open-weight vision-language flagship, released May 29, 2026. A 198B MoE activating ~11B per token, served over 400 tokens/second under Apache 2.0.

ERNIE 5.1

Baidu's closed-weight flagship, released May 9, 2026. A sparse Mixture-of-Experts model with a 128K context — the first ERNIE whose weights stay unpublished.