ScreenSpot
Qwen3-VL-235B-A22B 95.4% leads since September 2025 Qwen3-VL-235B-A22B 95.4% leads since September 2025 Model locates a described UI element in mobile, desktop and web screenshots. Scored by click-point accuracy.
Largely superseded by ScreenSpot-Pro for frontier comparisons.
- Made by
- Nanjing University (SeeClick)
- Released
- January 2024
- Kind
- Finding things on screen
- Level
- Everyday
- Scored
- Checked automatically
- Answers
- Short answer
- Tasks
- All public
- Last in a launch post
- December 2025, dropped after it saturated
- Last model added
- April 2026
- Scores refreshed
- November 12th, 2026
Qwen3-VL-235B-A22B leads with 95.4% among models released by April 2026
Best score per model as labs report them in their own cards. Labs pick their own settings, so these compare less cleanly than independent runs.
CogAgent leads with 47.4% across all independent runs
Best score per model across the official leaderboard, Epoch AI’s own runs and any board Epoch AI copied. Each bar and row names its source.
Qwen3-VL-235B-A22B leads with 95.4%, claims included
Best score per model from every source, including labs’ own numbers and claims Epoch AI collected from reports and papers. Labs pick their own settings, so a claim can beat an independent run of the same model.
The newest model with a score here came out in April 2026, and nobody has reported a newer one. Among the 20 highest-ranked models on the capability index, these have no score yet:
- Claude Fable 5.2
-
Gemini 4 Carbon - GPT-6 Astra
- Claude Opus 5.5
- and 16 more
- 1 Qwen3-VL-235B-A22B 95.4%
- 2
Amazon Nova Act 90.9% - 3
InternVL3.5-241B-A28B 89.8% - 4
Nemotron 3 Nano Omni 30B-A3B 89.3% - 5 Nova 2.0 Pro Preview 88.1%
- 6 Amazon Nova Premier 87.8%
- 7 Qwen2.5-VL-72B 87.1%
- 8
Fara-7B 86.7% - 9 Nova 2 Omni Preview 85.4%
- 10 Nova 2 Lite 83.3%
- 11 Nemotron Nano 12B v2 VL 40.1%
Source: Labs’ model cards
- 1 CogAgent 47.4%
Source: Official ScreenSpot leaderboard
- 1 Qwen3-VL-235B-A22B 95.4%
- 2 Amazon Nova Act 90.9%
- 3 InternVL3.5-241B-A28B 89.8%
- 4 Nemotron 3 Nano Omni 30B-A3B 89.3%
- 5 Nova 2.0 Pro Preview 88.1%
- 6 Amazon Nova Premier 87.8%
- 7 Qwen2.5-VL-72B 87.1%
- 8 Fara-7B 86.7%
- 9 Nova 2 Omni Preview 85.4%
- 10 Nova 2 Lite 83.3%
- 11 CogAgent 47.4%
- 12 Nemotron Nano 12B v2 VL 40.1%
Sources: Labs’ model cards and official ScreenSpot leaderboard
| # | Model | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | Qwen3-VL-235B-A22B | 95.4% | Sep 2025 | Lab model card | |
| 2 | Amazon Nova Act | 90.9% | Mar 2025 | Lab model card | |
| 3 | InternVL3.5-241B-A28B | 89.8% | Aug 2025 | Lab model card | |
| 4 | Nemotron 3 Nano Omni 30B-A3B | 89.3% | Apr 2026 | Lab model card | |
| 5 | Nova 2.0 Pro Preview | 88.1% | Dec 2025 | Lab model card | |
| 6 | Amazon Nova Premier | 87.8% | Apr 2025 | Lab model card | |
| 7 | Qwen2.5-VL-72B | 87.1% | Jan 2025 | Lab model card | |
| 8 | Fara-7B | 86.7% | Nov 2025 | Lab model card | |
| 9 | Nova 2 Omni Preview | 85.4% | Dec 2025 | Lab model card | |
| 10 | Nova 2 Lite | 83.3% | Dec 2025 | Lab model card | |
| 11 | Nemotron Nano 12B v2 VL | 40.1% | Nov 2025 | Lab model card |
| # | Model | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | CogAgent | 47.4% | Dec 2023 | Official leaderboard |
| # | Model | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | Qwen3-VL-235B-A22B | 95.4% | Sep 2025 | Lab model card | |
| 2 | Amazon Nova Act | 90.9% | Mar 2025 | Lab model card | |
| 3 | InternVL3.5-241B-A28B | 89.8% | Aug 2025 | Lab model card | |
| 4 | Nemotron 3 Nano Omni 30B-A3B | 89.3% | Apr 2026 | Lab model card | |
| 5 | Nova 2.0 Pro Preview | 88.1% | Dec 2025 | Lab model card | |
| 6 | Amazon Nova Premier | 87.8% | Apr 2025 | Lab model card | |
| 7 | Qwen2.5-VL-72B | 87.1% | Jan 2025 | Lab model card | |
| 8 | Fara-7B | 86.7% | Nov 2025 | Lab model card | |
| 9 | Nova 2 Omni Preview | 85.4% | Dec 2025 | Lab model card | |
| 10 | Nova 2 Lite | 83.3% | Dec 2025 | Lab model card | |
| 11 | CogAgent | 47.4% | Dec 2023 | Official leaderboard | |
| 12 | Nemotron Nano 12B v2 VL | 40.1% | Nov 2025 | Lab model card |
The benchmark's own leaderboard has 6 entries, 1 of them models we track. Switch the scores above to rank them together.
3 models have held number one in labs’ own numbers since its release in January 2024 Number one over time 4 models have held number one with claims included since its release in January 2024
Lab claims count from the day the card came out, independent runs from the model’s release.
Source: Labs’ model cards
Qwen’s own number for Qwen2.5-VL-72B passed 85% in January 2025
All the scores we have, one dot each, so a model tested at several settings appears more than once.
Sources: Official ScreenSpot leaderboard and labs’ model cards
Labs headline it less often now, last in December 2025
Launch posts from every lab that has listed it, one tick per post. Tall colored ticks include it, and short gray ones leave it out.
| Date | Launch post | Score | As the post writes it |
|---|---|---|---|
| Dec 2, 2025 | Nova 2 Omni Preview | 85.4 | ScreenSpot point accuracy |
| Dec 2, 2025 | Nova 2 Lite | 83.3 | ScreenSpot point accuracy |
| Dec 2, 2025 | Nova 2.0 Pro Preview | 88.1 | ScreenSpot point accuracy |
| Sep 23, 2025 | Qwen3-VL-235B-A22B | 95.4 | ScreenSpot |
| Aug 25, 2025 | InternVL3.5-241B-A28B | 89.8 | ScreenSpot |
| Apr 30, 2025 | Amazon Nova Premier | 87.8 | Mean of ScreenSpot Web Text 91.7 / Web Icon 84.0 |
| Mar 31, 2025 | Amazon Nova Act | 90.9 | Mean of ScreenSpot Web Text 0.939 / ScreenSpot Web Icon 0.879 |
| Jan 28, 2025 | Qwen2.5-VL-72B | 87.1 | ScreenSpot |
9 system cards report it, from 5 labs
Besides launch posts, we read the system cards, model cards and technical reports labs publish with a release. They report far more, including safety and frontier-risk tests.
| Date | Document | Score | As the document writes it | Section |
|---|---|---|---|---|
| Apr 27, 2026 | Nemotron 3 Nano Omni 30B-A3B technical report | 89.3 | ScreenSpot, reasoning on, 90.3 reasoning off (Table 7) | Capabilities |
| Dec 2, 2025 | Nova 2 Omni Preview technical report | 85.4 | ScreenSpot point accuracy (Table 6) | Capabilities |
| Dec 2, 2025 | Nova 2 Lite technical report | 83.3 | ScreenSpot point accuracy (Table 3) | Capabilities |
| Dec 2, 2025 | Nova 2.0 Pro Preview technical report | 88.1 | ScreenSpot point accuracy (Table 3) | Capabilities |
| Nov 24, 2025 | Fara-7B technical report | 86.7 | ScreenSpot-V1 (Table 13a) | Capabilities |
| Nov 6, 2025 | Nemotron Nano 12B v2 VL technical report | 40.1 | ScreenSpot, reasoning on (Table 3). Reasoning off 39.4 | Capabilities |
| Aug 25, 2025 | InternVL3.5-241B-A28B technical report | 89.8 | ScreenSpot (Table 2, Table 10) | Capabilities |
| Apr 30, 2025 | Amazon Nova Premier technical report | ScreenSpot Web Text 91.7 and Web Icon 84.0, no overall (Table 1) | Capabilities | |
| Feb 19, 2025 | Qwen2.5-VL-72B technical report | 87.1 | ScreenSpot (Table 9) | Capabilities |
Versions
We switch the top row to each new version when it comes out. Open it for one row per version. An old version stays in play while people still test it, so whoever first takes it past 85% gets credit for closing it out.
Source: Labs’ model cards
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 5.5 81.8%
- 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% - MobileWorld In use Agents complete 201 tasks across 20 mobile apps, including tasks that need user clarification or MCP tools. Backend checks verify success. GPT-5 55.3%