LithoBench - Practical Application (Open-Ended): leaderboard

LithoBench Practical Application (Open-Ended) scores vision-language models on 1,200 open-ended questions about practical assessments such as mapping, engineering and resource implications in GF-2 satellite image patches (LithoBench, 2026). The OPEN_J score, from 0 to 10 and higher is better, blends ROUGE-L and BLEU-4 overlap with the reference answer and a GPT-5.4 judge rating.

Metric: OPEN_J score (0-10) = 0.4 x ROUGE-L x 10 + 0.2 x BLEU-4 x 10 + 0.4 x the GPT-5.4 judge's overall score (0-10), comparing the answer with the reference answer, on the 1,200 open-ended practical application questions of LithoBench (practical assessments such as mapping, engineering and resource implications in GF-2 remote-sensing image patches); temperature 0, 1,024 output tokens; higher is better. Source: arxiv.org. Saturation forecast: Around 2029. 11 models tracked. Updated 7 Oct 2026.

Top models

#ModelScoreOverall rank
1InternVL3.5-241B-A28B3.96#733
2Kimi K2.53.95#266
3Intern-S1-Pro3.91#585
4GLM-4.6V3.86#710
5Qwen 3.5 397B A17B3.86#264
6Grok 43.69#319
7Gemini 3 Pro3.49#145
8InternVL3.5-8B-HF3.3
9Qwen 3.5 9B2.88#828
10Claude Sonnet 4.62.82#173
11Qwen 3.5 27B1.84#375

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

Interactive version: theaggregate.ai/benchmark?slug=lithobench-practical-application-open-ended · How It Works · Data refreshed daily, snapshot 2026-11-21.