Benchmark Almanac

AgentCLUE-Mobile (offline)

Nebula-GUI-V2 92.3 tops its leaderboard GLM-4.5V 88.4 leads the models we track
Not in model cards

Agents select actions in an offline mobile interface.

SuperCLUE reports scores in a public workbook.

Made by
SuperCLUE
Kind
Phones
Level
Everyday
Scored
Settled by events
Answers
Agent task
Measures
Capability
Language
Chinese
In and out
Text, Screens → Actions
Interaction
Single turn
Environment
Phone apps
Refresh
One fixed set
Run by
A host
Based on
AgentCLUE-Mobile
In views
Computer use
Strongest model tested
Gemini 2.5 Pro Jun 2025 #97 on the capability index
Last model added
August 2025
Scores refreshed
November 20th, 2026
12
Entries on its own leaderboard, 9 models we track

Nebula-GUI-V2 tops the leaderboard with 92.3 GLM-4.5V leads with 88.4 among models released by August 2025

Best score per entry from the benchmark's own leaderboard, read on November 20th, 2026. Where a model ran with several settings or harnesses, we keep its best. Names in gray aren't in our model list: fine-tunes, agent systems, products and older models. They don't count toward leaders. The best model we track is GLM-4.5V, at 88.4. This view shows only models we track, each at its best.

The newest model with a score here came out in August 2025, and nobody has tested a newer one. Among the 20 highest-ranked models on the capability index, these have no score yet:

  1. 1 Nebula-GUI-V2 API 92.3
  2. 2 Doubao-Seed-1.6-thinking-250715 API 89.9
  3. 3 GLM-4.5V API 88.4
  4. 4 MiMo-VL-7B-RL-2508 模型 84.4
  5. 5 Qwen2.5-VL-7B 模型 72.3
  6. 6 UI-TARS-1.5 API 66.2
  7. 7 GUI-Owl-7B 模型 64.2
  8. 8 Gemini 2.5 Pro Jun 2025, API 60.9
  9. 9 Qwen2.5-VL-3B 模型 57.9
  10. 10 AgentCPM-GUI 模型 54.4
  11. 11 MiniCPM-V 4.5 模型 39.1
  12. 12 Gemma 3 4B API 8.94

Source: Official AgentCLUE-Mobile (offline) leaderboard

#EntryScoreSettingDateSource
1 Nebula-GUI-V2 92.3 API Official leaderboard
2 Doubao-Seed-1.6-thinking-250715 89.9 API Official leaderboard
3 GLM-4.5V 88.4 API Aug 2025 Official leaderboard
4 MiMo-VL-7B-RL-2508 84.4 模型 Aug 2025 Official leaderboard
5 Qwen2.5-VL-7B 72.3 模型 Jan 2025 Official leaderboard
6 UI-TARS-1.5 66.2 API Apr 2025 Official leaderboard
7 GUI-Owl-7B 64.2 模型 Aug 2025 Official leaderboard
8 Gemini 2.5 Pro Jun 2025 60.9 API Jun 2025 Official leaderboard
9 Qwen2.5-VL-3B 57.9 模型 Jan 2025 Official leaderboard
10 AgentCPM-GUI 54.4 模型 Official leaderboard
11 MiniCPM-V 4.5 39.1 模型 Aug 2025 Official leaderboard
12 Gemma 3 4B 8.94 API Mar 2025 Official leaderboard
#ModelScoreSettingDateSource
1 GLM-4.5V 88.4 API Aug 2025 Official leaderboard
2 MiMo-VL-7B-RL-2508 84.4 模型 Aug 2025 Official leaderboard
3 Qwen2.5-VL-7B 72.3 模型 Jan 2025 Official leaderboard
4 UI-TARS-1.5 66.2 API Apr 2025 Official leaderboard
5 GUI-Owl-7B 64.2 模型 Aug 2025 Official leaderboard
6 Gemini 2.5 Pro Jun 2025 60.9 API Jun 2025 Official leaderboard
7 Qwen2.5-VL-3B 57.9 模型 Jan 2025 Official leaderboard
8 MiniCPM-V 4.5 39.1 模型 Aug 2025 Official leaderboard
9 Gemma 3 4B 8.94 API Mar 2025 Official leaderboard

Rankings by part

We read 20 parts from the benchmark’s own leaderboard in 3 groups: 不同指令类型得分榜单, 综合指令执行质量评分榜单 and 应用场景榜单. Each ranks models on one part of the test. GLM-4.5V leads 14 of them, and 3 other models lead the rest. Arrows show how far a model moved from its overall rank among the models we track.

UI元素定位平均分

9 models we track have a score here.

  1. 1 Qwen2.5-VL-7B ↑2 94.9
  2. 2 GLM-4.5V ↓1 94.8
  3. 3 GUI-Owl-7B ↑2 94.2
  4. 4 Qwen2.5-VL-3B ↑3 89.9
  5. 5 MiMo-VL-7B-RL-2508 ↓3 88.0
  6. 6 UI-TARS-1.5 ↓2 79.9
  7. 7 MiniCPM-V 4.5 ↑1 74.0
  8. 8 Gemini 2.5 Pro Jun 2025 ↓2 58.0
  9. 9 Gemma 3 4B 8.00

3 models have held number one on the official leaderboard since 2023

Lab claims count from the day the card came out, independent runs from the model’s release.

Qwen Xiaomi Corp Z.ai
Own leaderboard Not tested since Aug 2025
GLM-4.5V 88.4

Source: Official AgentCLUE-Mobile (offline) leaderboard

Scores over time

All the scores we have, one dot each, so a model tested at several settings appears more than once.

Plot with other benchmarks
On the benchmark’s own leaderboard
0.00 20.0 40.0 60.0 80.0 100.0 2025 0.00 20.0 40.0 60.0 80.0 100.0 2025

Source: Official AgentCLUE-Mobile (offline) leaderboard

No lab model card reports it yet

We read the headline results of 540 launch posts from 57 labs, and none of them list this benchmark.

More in Computer and browser use

All 119 in Computer and browser use
  • BrowseComp In use Hard-to-find facts that need persistent web browsing. Short verifiable answers graded for exact correctness. Claude Fable 5 93.3%
  • ScreenSpot-Pro In use GUI grounding on high-resolution professional-software screenshots. The model must locate the target UI element. GPT-6 Astra 92.7%
  • AndroidWorld In use Agents complete 116 tasks across 20 real Android apps in an emulator. Each task checks the device's system state to judge success. GPT-6 Astra 100%
  • OSWorld 2.1 In use Agents complete 108 long-horizon workflows across 31 self-hosted websites and desktop apps, averaging 318 tool calls. Scored by binary completion and checkpoint partial credit. Claude Opus 5 44.3%
  • WebArena-Verified In use Agents complete 812 audited WebArena tasks on self-hosted websites. Deterministic evaluators check the agent's answer and captured network traces. Muse Spark 1.1 69%
  • OSWorld 2.1 (offline subset) In use Agents complete the 82 OSWorld 2.1 tasks on the official offline list, with no internet access. Scored by checkpoint partial credit and strict pass rate. Claude Fable 5.2 89.3%

Data audit

7 of 9 checks pass, 2 warnings. Look first at the Task access check. We don’t know whether the tasks are public. Every benchmark

Catalog record 1 warning 1 pass
  • Warning Task access We don’t know whether the tasks are public.
  • Pass Confidence Checked against a primary source.
Scores 1 warning 2 pass 2 notes
  • Pass Independent runs Scores from the official leaderboard (12 entries).
  • Pass Models scored 9 models we track, plus 3 other entries on its leaderboard.
  • Warning Newest score The newest came 15 months ago, in August 2025, for MiniCPM-V 4.5.
  • Note Frontier models None of the top 10 has a score. The strongest tested is Gemini 2.5 Pro (#97).
  • Note Headroom Scored as a plain number, with no fixed ceiling.
Leaderboard import 3 pass
  • Pass Importer We read the board on superclueai.com: 9 rows for models we track and 3 other entries at the last read.
  • Pass Last read November 20th, 2026.
  • Pass Read worked The last read worked.
Consistency 1 pass 1 note
  • Note Lab claims No model has both a lab number and an independent run.
  • Pass Known issues No published audit found problems.

How we check