LongShOTBench - Core Perception: leaderboard

LongShOTBench (arXiv:2512.16978) Core Perception (CP) category breakdown: weighted criterion-level rubric score (0-100%) on fine-grained perceptual recognition across long omni-modal videos; higher is better.

Metric: Weighted Rubric Score (%). Source: longshot.cvmbzuai.com. Saturation forecast: Around 2030. 102 models tracked. Updated 21 Nov 2026.

Top models

#ModelScoreOverall rank
1Qwen 3 Omni 30B A3B (Thinking)51.23
2Gemini 3.1 Pro (Preview)44.67#99
3Voxtral-Mini-3B-250741.37#3682
4Vero-Qwen3I-8B41.08
5Qwen 3 VL 235B A22B (Thinking)40.5#459
6Voxtral-Small-24B-250739.66#1246
7Qwen 3 VL 32B (Thinking)39.39#573
8Kimi K2.638.16#192
9Qwen 3 VL 235B A22B Instruct36.64#532
10Qwen 3 VL 30B A3B (Thinking)36.52#719
11Gemma 3 27B (IT)35.43#1324
12Qwen 3 VL 32B Instruct35.34#606
13Qwen 3 VL 8B (Thinking)35.11#910
14Intern-S1-mini34.47#1178
15Gemma 3 12B (IT)32.9#1887

No result here: #1 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.

Interactive version: theaggregate.ai/benchmark?slug=longshotbench-core-perception · How It Works · Data refreshed daily, snapshot 2026-11-21.