ZMIME

WeMM-Video-35B-A3B vs Inkling 2 Mini

Compare documented capabilities and access. A dash means a comparable figure is unavailable.

Model specifications
SpecificationWeMM-Video-35B-A3BTencentInkling 2 MiniThinking Machines
TaskAudio-visual understandingMultimodal reasoning and coding
Published capabilitiesAnswers questions in Chinese or English about video frames and their soundtrack. Supports direct responses and a thinking mode.Reuses its middle layers three times per token. Thinking Machines reports 48.2% on DeepSWE v1.1.
Model size35B total / 3B active MoE backbone26.4B total / 3.4B active parameters
AccessPublic BF16 checkpoint and quantized releasesDownloadable weights and Tinker inference and customization
ResultText responses, including descriptions and speech transcripts—
Operating limits—Native mixed-precision checkpoint occupies approximately 20 GB.
SourcesModel weights · Additional documentationModel weights · Additional documentation
InputsText, Images, Audio, VideoText, Images
WeightsOpen weights · Apache-2.0Open weights · Apache-2.0
ReleasedNov 20, 2026Nov 20, 2026