GLM-5V-Turbo

Leaderboards

Where GLM-5V-Turbo sits among the top 20 in each category.

By task

Scored from votes on each kind of prompt alone, on the same 1,000-point scale.

Test results

Every test GLM-5V-Turbo has a published result on, out of 100%, with its rank among the models that took it.

• OSWorld

  #7 / 21

  OSWorld: The first-of-its-kind scalable, real computer environment for multimodal agents, supporting task setup, execution-based evaluation, and interactive learning across Ubuntu, Windows, and macOS with 369 computer tasks involving real web and desktop applications, OS file I/O, and multi-application workflows

  62.3%

  Self-reported

• FACTS Grounding

  #17 / 20

  A benchmark evaluating language models' ability to generate factually accurate and well-grounded responses based on long-form input context, comprising 1,719 examples with documents up to 32k tokens requiring detailed responses that are fully grounded in provided documents

  58.6%

  Self-reported

• SimpleVQA

  #3 / 17

  SimpleVQA is a visual question answering benchmark focused on simple queries.

  0.8%

  Self-reported

• Claw-Eval

  #2 / 14

  Claw-Eval tests real-world agentic task completion across complex multi-step scenarios, evaluating a model's ability to use tools, navigate environments, and complete end-to-end tasks autonomously.

  75.0%

  Self-reported

• AndroidWorld

  #6 / 8

  AndroidWorld evaluates an agent's ability to operate in real Android GUI environments, completing multi-step tasks by perceiving screen content and executing touch/type actions.

  75.7%

  Self-reported

• PinchBench

  #6 / 7

  PinchBench evaluates coding agents on real-world agentic coding tasks, measuring both best-case and average performance across complex software engineering scenarios.

  80.7%

  Self-reported

• V*

  #7 / 7

  A visual reasoning benchmark evaluating multimodal inference under challenging spatial and grounded tasks.

  89.0%

  Self-reported

• Vision2Web

  #5 / 5

  Vision2Web evaluates multimodal models on converting visual designs and screenshots into functional web pages, measuring end-to-end design-to-code capability.

  31.0%

  Self-reported

• ZClawBench

  #2 / 4

  ZClawBench evaluates Claw-style agent task execution quality, measuring a model's ability to autonomously complete complex multi-step coding tasks in real-world environments.

  57.6%

  Self-reported

• Design2Code

  #1 / 2

  Design2Code evaluates the ability to generate code (HTML/CSS/JS) from visual designs.

  0.9%

  Unverified

• MMSearch-Plus

  #2 / 2

  MMSearch-Plus is an extended variant of MMSearch with harder multimodal search and retrieval tasks requiring deeper reasoning over visual and textual search results.

  30.0%

  Unverified

• BrowseComp-VL

  #1 / 1

  BrowseComp-VL is the vision-language variant of BrowseComp, evaluating multimodal models on web browsing comprehension tasks that require processing visual web page content alongside text.

  51.9%

  Unverified

• CC-Bench-V2 Backend

  #1 / 1

  CC-Bench-V2 Backend evaluates coding agents on backend development tasks, measuring practical engineering ability to implement server-side logic, APIs, and system components.

  22.8%

  Unverified

• CC-Bench-V2 Frontend

  #1 / 1

  CC-Bench-V2 Frontend evaluates coding agents on frontend development tasks, measuring ability to build UI components, handle styling, and implement client-side logic.

  68.4%

  Unverified

• CC-Bench-V2 Repo Exploration

  #1 / 1

  CC-Bench-V2 Repo Exploration evaluates coding agents on repository-level understanding and navigation, measuring ability to explore, comprehend, and work across entire codebases.

  72.2%

  Unverified

• Flame-VLM-Code

  #1 / 1

  Flame-VLM-Code evaluates multimodal models on visual code generation tasks, measuring ability to generate code from visual inputs such as UI mockups and design specifications.

  93.8%

  Unverified

• ImageMining

  #1 / 1

  ImageMining evaluates multimodal models on extracting structured information from images using tool use, measuring ability to combine visual understanding with tool-based retrieval and analysis.

  30.7%

  Unverified

• MMSearch

  #1 / 1

  MMSearch evaluates multimodal models on search-based retrieval and question answering tasks that require processing both visual and textual information from search results.

  72.9%

  Unverified

• WebVoyager

  #1 / 1

  WebVoyager evaluates an agent's ability to navigate and complete tasks on real websites by perceiving page screenshots and executing browser actions.

  88.5%

  Unverified