KnowU-Bench
Agents use Android apps such as Mattermost and Taodian. They infer preferences, ask for missing details, and decide when to act or stay silent.
The leaderboard scores eight models on 42 general, 86 personalized, and 64 proactive tasks. The authors combine state checks with model judgments.
- Made by
- Zhejiang University, Apple, and Tencent
- Released
- April 2026
- Kind
- Phones
- Level
- Everyday
- Scored
- Checked automatically
- Answers
- Agent task
- Tasks
- All public
- Measures
- Capability
- Language
- English
- Interaction
- Agentic
- Environment
- Phone apps
- Items from
- Generated
- Refresh
- One fixed set
- Run by
- Anyone
- In views
- Computer use, Agents
- Strongest model tested
-
Gemini 3.1 Pro #40 on the capability index
- Last model added
- February 2026
- Scores refreshed
- November 20th, 2026
Claude Sonnet 4.6 leads with 60.4% among models released by February 2026 Claude Sonnet 4.6 leads with 60.4% among models released by February 2026
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. This view shows only models we track, each at its best.
The newest model with a score here came out in February 2026, and nobody has tested a newer one. Among the 20 highest-ranked models on the capability index, these have no score yet:
- GPT-6.2 Astra
- Claude Fable 5.2
-
Gemini 4 Carbon
- GPT-6 Astra
- and 16 more
- 1 Claude Sonnet 4.6 End-to-End, 50 steps 60.4%
- 2
Seed2.0 Pro End-to-End, 50 steps 51.6%
- 3
Gemini 3.1 Pro End-to-End, 50 steps 44.3%
- 4
Qwen3.5 397B-A17B End-to-End, 50 steps 37.5%
- 5
Qwen3.5-9B End-to-End, 50 steps 33.3%
- 6
Qwen3-VL-32B-Instruct End-to-End, 50 steps 29.3%
- 7 UI-Venus-1.5-8B End-to-End, 50 steps 26%
- 8 Qwen3-VL-8B End-to-End, 50 steps 21.9%
Source: Official KnowU-Bench leaderboard
Source: Official KnowU-Bench leaderboard
| # | Entry | Score | Setting | Released | Run on | Source |
|---|---|---|---|---|---|---|
| 1 | Claude Sonnet 4.6 | 60.4% | End-to-End, 50 steps | Feb 2026 | Apr 2026 | Official leaderboard |
| 2 | | 51.6% | End-to-End, 50 steps | Feb 2026 | Apr 2026 | Official leaderboard |
| 3 | | 44.3% | End-to-End, 50 steps | Feb 2026 | Apr 2026 | Official leaderboard |
| 4 | | 37.5% | End-to-End, 50 steps | Feb 2026 | Apr 2026 | Official leaderboard |
| 5 | | 33.3% | End-to-End, 50 steps | Feb 2026 | Apr 2026 | Official leaderboard |
| 6 | | 29.3% | End-to-End, 50 steps | Oct 2025 | Apr 2026 | Official leaderboard |
| 7 | UI-Venus-1.5-8B | 26% | End-to-End, 50 steps | Apr 2026 | Official leaderboard | |
| 8 | Qwen3-VL-8B | 21.9% | End-to-End, 50 steps | Apr 2026 | Official leaderboard |
| # | Model | Score | Setting | Released | Run on | Source |
|---|---|---|---|---|---|---|
| 1 | Claude Sonnet 4.6 | 60.4% | End-to-End, 50 steps | Feb 2026 | Apr 2026 | Official leaderboard |
| 2 | | 51.6% | End-to-End, 50 steps | Feb 2026 | Apr 2026 | Official leaderboard |
| 3 | | 44.3% | End-to-End, 50 steps | Feb 2026 | Apr 2026 | Official leaderboard |
| 4 | | 37.5% | End-to-End, 50 steps | Feb 2026 | Apr 2026 | Official leaderboard |
| 5 | | 33.3% | End-to-End, 50 steps | Feb 2026 | Apr 2026 | Official leaderboard |
| 6 | | 29.3% | End-to-End, 50 steps | Oct 2025 | Apr 2026 | Official leaderboard |
Rankings by part
We read 3 parts from the benchmark’s own leaderboard: General, Personalized and Proactive. Each ranks models on one part of the test. Claude Sonnet 4.6 leads every one of them. Arrows show how far a model moved from its overall rank among the models we track.
General
6 models we track have a score here.
- 1 Claude Sonnet 4.6 81%
- 2
Seed2.0 Pro 78.6%
- 3
Gemini 3.1 Pro 78.6%
- 4
Qwen3-VL-32B-Instruct ↑2 47.6%
- 5
Qwen3.5 397B-A17B ↓1 47.6%
- 6
Qwen3.5-9B ↓1 42.9%
Personalized
6 models we track have a score here.
- 1 Claude Sonnet 4.6 44.2%
- 2
Seed2.0 Pro 30.2%
- 3
Gemini 3.1 Pro 27.9%
- 4
Qwen3.5 397B-A17B 13.9%
- 5
Qwen3-VL-32B-Instruct ↑1 10.6%
- 6
Qwen3.5-9B ↓1 4.70%
Proactive
6 models we track have a score here.
- 1 Claude Sonnet 4.6 68.8%
- 2
Qwen3.5-9B ↑3 65.6%
- 3
Qwen3.5 397B-A17B ↑1 62.5%
- 4
Seed2.0 Pro ↓2 62.5%
- 5
Gemini 3.1 Pro ↓2 43.8%
- 6
Qwen3-VL-32B-Instruct 42.2%
Number one over time
Lab claims count from the day the card came out, independent runs from the model’s release.
Source: Official KnowU-Bench leaderboard
Scores over time
All the scores we have, one dot each, so a model tested at several settings appears more than once.
Source: Official KnowU-Bench 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
- 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
9 of 10 checks pass, 1 warning. Look first at the Newest score check. The newest came 9 months ago, in February 2026, for Qwen3.5-9B. Every benchmark
Catalog record 2 pass
- Pass Task access All tasks are public.
- Pass Confidence Checked against a primary source.
Scores 1 warning 3 pass 1 note
- Pass Independent runs Scores from the official leaderboard (8 entries).
- Pass Models scored 6 models we track, plus 2 other entries on its leaderboard.
- Warning Newest score The newest came 9 months ago, in February 2026, for Qwen3.5-9B.
- Note Frontier models None of the top 10 has a score. The strongest tested is Gemini 3.1 Pro (#40).
- Pass Headroom The best is 60.4%, under 85%, so it still separates models.
Leaderboard import 3 pass
- Pass Importer We read the board on zjureal.com: 6 rows for models we track and 2 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.