Inkling 2 Small Model Card - ScreenSpot Pro - Original: leaderboard
November 5 publisher model-card comparison. Inkling 2 Small is reported at effort 0.99 (xhigh), temperature 1.0 and a 1M-token maximum trajectory. Comparator effort, generation settings and dated snapshots are not specified. No rc17 or Vals identity is inferred. Original and Python-enabled variants remain separate. Images whose long side exceeds 4096 pixels are downsampled preserving aspect ratio; ground truth is rescaled. Smaller images are unchanged.
Metric: Published score (%). Source: thinkingmachines.ai. Saturation forecast: Rough model projection: around 2026. 5 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Inkling 2 Small (Benchmark Sweep Run: source=Inkling2SmallCard; original; effort=0.99 xhigh; temperature=1; trajectory=1M tokens) | 83.8 | |
| 2 | GPT-6 Luna (Benchmark Sweep Run: source=Inkling2SmallCard; original; effort, temperature and snapshot unreported) | 81.1 | |
| 3 | Kimi K3 (Benchmark Sweep Run: source=Inkling2SmallCard; original; effort, temperature and snapshot unreported) | 73.8 | |
| 4 | MiMo V2.6 Pro (Benchmark Sweep Run: source=Inkling2SmallCard; original; effort, temperature and snapshot unreported) | 56.6 | |
| 5 | Mistral Large 4 (Benchmark Sweep Run: source=Inkling2SmallCard; original; effort, temperature and snapshot unreported) | 42.3 |
Interactive version: theaggregate.ai/benchmark?slug=inkling-2-small-model-card-screenspot-pro-original · How It Works · Data refreshed daily, snapshot 2026-11-06.