WeMM-Video Models by Tencent AI Lab

Tencent AI LabApache 2.0Open source
Tencent AI Lab releases · 18 in the last 12 months · this family litChangelog →
4 models2026Up to 256K ctx

Last refreshed 2026-11-20. Next refresh: weekly.

Details

ResearcherTencent AI Lab
LicenseApache 2.0OSI-approved
Commercial usePermitted
Models4
Released2026
Max context256K

Capabilities

VisionAll models
MultimodalAll models
ReasoningAll models

Links

Website

About

Tencent WeChat Vision’s bilingual audio-visual understanding and reasoning models.

Decision facts

Best fit
vision and multimodal workreasoningcoding
Capability starting point
WeMM-Video 2B with 256K context and reasoning and multimodal inputs
Lowest tracked input
Not tracked
Closest related family
Hunyuan Image

Current Variants

Use-when guidance is based on each model's tracked capabilities, context window, release date, and replacement status.

4 in view

Use when the workload needs multimodal, 256K context, and 2B parameters.

2026-11multimodal256K context2B parameters

Use when the workload needs multimodal, 256K context, and 4B parameters.

2026-11multimodal256K context4B parameters

Use when the workload needs multimodal, 256K context, and 9B parameters.

2026-11multimodal256K context9B parameters

Use when the workload needs multimodal, 256K context, and reasoning.

2026-11multimodal256K contextreasoning

Release Timeline

1 release group
2026-11
4 current
WeMM-Video 2B
multimodal256K context2B parameters
Current
WeMM-Video 35B-A3B
multimodal256K contextreasoning
Current
WeMM-Video 4B
multimodal256K context4B parameters
Current
WeMM-Video 9B
multimodal256K context9B parameters
Current

Specifications(4 models)

WeMM-Video model specifications comparison
ModelReleasedContextParametersVisionMultimodalReasoning
WeMM-Video 2B2026-11256K2BYesYesYes
WeMM-Video 4B2026-11256K4BYesYesYes
WeMM-Video 9B2026-11256K9BYesYesYes
WeMM-Video 35B-A3B2026-11256K35B (3B active)YesYesYes