The Open Weights
LatestModelsLeaderboardsCompanies
Subscribe
The Open Weights

The daily record of open-source AI. New model releases, leaderboards, and what's coming next — written for people who ship.

Refreshed every 12 hours

Discover

  • Latest releases
  • New today
  • Trending models

Browse

  • All models
  • Companies
  • Categories
  • Leaderboards

About

  • About
  • Editorial policy
  • Hardware estimates
  • RSS feed
  • llms.txt
  • Newsletter

© 2026 The Open Weights. An independent publication.

PrivacyTermsSMSAggregated by Claude · curated by humans.
LatestTencent27B
TencentVision-Language

Tencent's UI-Mate-27B targets desktop automation

A 27B vision-language model built to see and operate graphical interfaces, tested on OSWorld and WindowsAgentArena.

Aug 14, 2026
NotableOther
UI-Mate-27B

Tencent has published UI-Mate-27B, a 27-billion-parameter vision-language model designed to act as a computer-use agent — reading what's on screen and carrying out multi-step tasks across desktop applications. Unlike a general-purpose chat model, it is tuned specifically for the loop of perceiving an interface, deciding on an action, and executing it.

The model is a dense (non-mixture-of-experts) VLM, placing it in the mid-size tier where capability and deployability meet. Tencent points to evaluation on two well-known agent benchmarks, OSWorld and WindowsAgentArena, which measure how reliably a system can complete real tasks inside actual operating-system environments rather than in simplified sandboxes.

Why it matters

GUI agents are one of the more demanding frontiers in applied AI: a model has to ground language in pixels, track state across steps, and recover from mistakes. Open weights in this category are still relatively scarce, so a 27B model aimed squarely at desktop control gives researchers and builders something concrete to test and fine-tune.

  • Purpose-built for computer-use and GUI automation, not generic vision tasks
  • 27B dense parameters, released under a custom ("other") license
  • Benchmarked on OSWorld and WindowsAgentArena

As always with agent models, the practical questions — latency, reliability on unfamiliar apps, and license terms for commercial use — will determine how far UI-Mate-27B travels beyond the benchmark tables. The weights and details are available now on Hugging Face.

Sources

  • tencent/UI-Mate-27B

    Hugging Face

    Visit
OlderDeepSeek Releases V4-Pro-0813 With Open WeightsDeepSeek · Text / LLM · 3 months agoNewerStarDoc-AI Releases TeleOCR for Document ParsingStarDoc AI · Vision-Language · 3 months ago

Get the model

Hugging Face

Specs

Parameters27B
Context window262K tokens
Size54.7 GB
PrecisionBF16
ArchitectureQwen3_5ForConditionalGeneration
LicenseOTHER
Downloads529
Likes95

Can you run it?

Runs on a 24 GB GPU at 4-bit.

  • BF16 (as published)

    80 GB GPU (A100 / H100) · 128 GB Mac Studio

    57.8 GB
  • 8-bit

    48 GB GPU (RTX 6000) or 2×24 GB · 64 GB Mac

    32.2 GB
  • 4-bit

    24 GB GPU (RTX 3090 / 4090) · 32 GB Mac

    19.6 GB
Your machine
GGUF builds
  • KV cache sized for 8,192 tokens of context; longer prompts need proportionally more.

Based on Hugging Face weights metadata. Assumes an 8K context and ~1 GB runtime overhead; actual needs vary. How we estimate


Modalities

Vision-Language

The Weekly Weights

Every open release that mattered, one email a week.

0 comments

No comments yet. Be the first to weigh in.

More from Tencent

All Tencent releases →
Hunyuan Hy4-flash
Tencent/Vision-LanguageMulti-GPU server

Tencent releases Hunyuan Hy4-flash multimodal MoE

The lightweight, mixture-of-experts model adds vision and text instruction-following to Tencent's open Hunyuan lineup.

Nov 17, 2026
Hunyuan Hy4
Tencent/Text / LLMDatacenter

Tencent releases Hunyuan Hy4, an MoE vision-language model

The instruction-tuned, mixture-of-experts model handles both images and text under an open weights listing on Hugging Face.

Oct 29, 2026
Hy-Nomas
Tencent/Text → Video

Tencent Releases Hy-Nomas, an Interactive Video World Model

The open-weights model pairs text- and image-to-video generation with action-aware causal attention and explicit camera control.

Oct 21, 2026

More in Vision-Language

All Vision-Language →
SenseNova-Loop-9B
SenseTime/Vision-LanguageRuns on a laptop

SenseTime Releases Looped 9B Multimodal Model

The company's new SenseNova-Loop-9B pairs vision-language capabilities with a recurrent looped-transformer design.

Nov 17, 2026
Inkling 2 Mini
Thinkingmachines/Vision-LanguageConsumer GPU

Thinking Machines releases Inkling 2 Mini, a compact open MoE

The Apache-2.0 multimodal model uses a mixture-of-experts design aimed at running locally rather than in the cloud.

Nov 17, 2026
KORMo-Web-11B
KORMo VL/Vision-LanguageRuns on a laptop

KORMo-Web-11B targets Korean and English GUI agents

An 11-billion-parameter vision-language model built to read screens and drive web and desktop interfaces across two languages.

Nov 14, 2026