Tool tag: Open Weights

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.

Command A+

Cohere's first Mixture-of-Experts model, released May 2026. 218B total / 25B active parameters, vision input, reasoning, and agentic tool use under Apache 2.0.

Ling-2.6-1T

Ant Group's open-weight, non-thinking flagship: ~1T parameters with 63B active, a hybrid MLA and Linear Attention stack, a 1M-token context, and an MIT license.

Mistral Medium 3.5

Mistral AI's multimodal model for agentic and coding work: a dense 128B model with a 256K context window, released April 2026 under a modified MIT license.

DeepSeek V4

DeepSeek's open-weight V4 family: V4-Pro (1.6T/49B), V4-Flash (284B/13B) and the experimental V4-Flash-Vision-Exp. MIT weights, 1M context.

MiMo-V2.5-Pro

Xiaomi's open-weight flagship, released April 22, 2026. A 1.02-trillion-parameter MoE with 42B active, a 1M-token context, and an MIT license.

Gemma 4

Google's open-weight Gemma 4 family — Apache 2.0 models from a 2.3B-effective edge model to a 30.7B dense flagship, with QAT builds for local deployment.

MOSS-TTS

OpenMOSS's open-source speech synthesis model (Feb 6, 2026). An 8B autoregressive TTS system under Apache 2.0, with zero-shot voice cloning and 20 languages.

HunyuanImage 3.0

Tencent's 80B-parameter MoE image model, released September 28, 2025. Natively multimodal and autoregressive, under a restricted community license.