# High-level instruction following on AndroidControl Source: [High-level instruction following on AndroidControl]() Collection index: [llms.txt]() Data updated: 2026-03-24T12:34:38.309Z Evaluation of model performance on high-level instruction following tasks using the AndroidControl dataset, reporting Grounding and Step Accuracy metrics. A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage. | Scope | Recorded value | | --- | --- | | Task | High-level instruction following | | Dataset (including any stated split) | AndroidControl | | Metric | Step Accuracy | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 16 of 16 | | Coverage | All results currently returned for this benchmark | Evaluation settings may differ between results. The original papers and recorded configurations are needed to establish comparability. A shared metric name alone does not establish an identical protocol. ## Metrics - [Step Accuracy](): Higher is better; unit: PERCENT. - [Grounding](): Higher is better; unit: PERCENT. ## Recorded results | Method | Step Accuracy | Result date | Report paper | | --- | --- | --- | --- | | Falcon-UI-7B | 72.7 | Not recorded | 68fa7ba21e209131cea897a8 | | MobileIPL | 72.7 | Not recorded | 68fa7ba21e209131cea897a8 | | UI-Tars-7B | 72.5 | Not recorded | 68fa7ba21e209131cea897a8 | | OS-Atlas-7B | 71.2 | Not recorded | 68fa7ba21e209131cea897a8 | | MobileIPL-7B | 70 | Not recorded | 68fa7ba21e209131cea897a8 | | Qwen2-VL-7B(GRPO) | 69.8 | Not recorded | 68fa7ba21e209131cea897a8 | | Qwen2-VL-7B (SFT) | 69.1 | Not recorded | 68fa7ba21e209131cea897a8 | | Qwen2-VL-7B(GRPO) | 68.1 | Not recorded | 68fa7ba21e209131cea897a8 | | InternVL-2-4B | 66.7 | Not recorded | 68fa7ba21e209131cea897a8 | | Qwen2-VL-7B(SFT) | 61.4 | Not recorded | 68fa7ba21e209131cea897a8 | | OS-Atlas-7B | 60.7 | Not recorded | 68fa7ba21e209131cea897a8 | | PaLM 2S(full) | 58.7 | Not recorded | 68fa7ba21e209131cea897a8 | | PaLM 2S(LoRA) | 58.5 | Not recorded | 68fa7ba21e209131cea897a8 | | SphAgent-7B | 57.1 | Not recorded | 68fa7ba21e209131cea897a8 | | FedMobileAgent | 52.3 | Not recorded | 68fa7ba21e209131cea897a8 | | Aria-UI-7B | 10.2 | Not recorded | 68fa7ba21e209131cea897a8 | ### Falcon-UI-7B Step Accuracy: 72.7. Result ID: 69c2840147f34ce924aa7718. [Reporting paper]() Recorded evaluation configuration: {"Mode":"PF"}. This record does not assert independent reproduction. ### MobileIPL Step Accuracy: 72.7. Result ID: 69c2840147f34ce924aa771b. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Mode":"Ours"}. This record does not assert independent reproduction. ### UI-Tars-7B Step Accuracy: 72.5. Result ID: 69c2840147f34ce924aa7719. [Reporting paper]() Recorded evaluation configuration: {"Mode":"PF"}. This record does not assert independent reproduction. ### OS-Atlas-7B Step Accuracy: 71.2. Result ID: 69c2840147f34ce924aa7717. [Reporting paper]() Recorded evaluation configuration: {"Mode":"PF"}. This record does not assert independent reproduction. ### MobileIPL-7B Step Accuracy: 70. Result ID: 69c2840147f34ce924aa772b. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Mode":"IPL"}. This record does not assert independent reproduction. ### Qwen2-VL-7B(GRPO) Step Accuracy: 69.8. Result ID: 69c2840147f34ce924aa771a. [Reporting paper]() Recorded evaluation configuration: {"Mode":"RL"}. This record does not assert independent reproduction. ### Qwen2-VL-7B (SFT) Step Accuracy: 69.1. Result ID: 69c2840147f34ce924aa7716. [Reporting paper]() Recorded evaluation configuration: {"Mode":"FT"}. This record does not assert independent reproduction. ### Qwen2-VL-7B(GRPO) Step Accuracy: 68.1. Result ID: 69c2840147f34ce924aa772a. [Reporting paper]() Recorded evaluation configuration: {"Mode":"RL"}. This record does not assert independent reproduction. ### InternVL-2-4B Step Accuracy: 66.7. Result ID: 69c2840147f34ce924aa7715. [Reporting paper]() Recorded evaluation configuration: {"Mode":"FT"}. This record does not assert independent reproduction. ### Qwen2-VL-7B(SFT) Step Accuracy: 61.4. Result ID: 69c2840147f34ce924aa7726. [Reporting paper]() Recorded evaluation configuration: {"Mode":"FT"}. This record does not assert independent reproduction. ### OS-Atlas-7B Step Accuracy: 60.7. Result ID: 69c2840147f34ce924aa7729. [Reporting paper]() Recorded evaluation configuration: {"Mode":"PF"}. This record does not assert independent reproduction. ### PaLM 2S(full) Step Accuracy: 58.7. Result ID: 69c2840147f34ce924aa7724. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Mode":"FT"}. This record does not assert independent reproduction. ### PaLM 2S(LoRA) Step Accuracy: 58.5. Result ID: 69c2840147f34ce924aa7725. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Mode":"FT"}. This record does not assert independent reproduction. ### SphAgent-7B Step Accuracy: 57.1. Result ID: 69c2840147f34ce924aa7728. [Reporting paper]() Recorded evaluation configuration: {"Mode":"PF"}. This record does not assert independent reproduction. ### FedMobileAgent Step Accuracy: 52.3. Result ID: 69c2840147f34ce924aa7727. [Reporting paper]() Recorded evaluation configuration: {"Mode":"PF"}. This record does not assert independent reproduction. ### Aria-UI-7B Step Accuracy: 10.2. Result ID: 69c2840147f34ce924aa7714. [Reporting paper]() Recorded evaluation configuration: {"Mode":"FT"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.