FIKA-Bench: leaderboard

FIKA-Bench (2026) asks for the exact fine-grained identity of a product, species, vehicle, landmark or cultural object in an image, with human-verified evidence for every answer and items kept only when strong closed-book models failed. Closed-book models answer from the image and question alone. Strict accuracy in percent from an LLM judge that accepts only answers at the required granularity; higher is better.

Metric: Overall strict answer accuracy (%) on FIKA-Bench, open-ended fine-grained recognition questions (311 images from decontaminated public fine-grained datasets and real-life photos, each answer backed by human-verified evidence URLs, kept after strong closed-book models failed); a text-only LLM judge counts only answers at the required granularity as correct, closed-book models (no tools, image and question only); higher is better. Source: arxiv.org. Saturation forecast: Around August 2027. 10 models tracked. Updated 7 Oct 2026.

Top models

#ModelScoreOverall rank
1Kimi K2.625.1#192
2Gemini 3.1 Flash Lite20.6#371
3Qwen 3.5 397B A17B18.3#264
4GPT-5 Mini17.4#364
5GLM-5V Turbo15.8#360
6Qwen 3 VL 235B A22B14.8#485
7Fine-R1-7B11.6
8Qwen 3.5 9B9.3#828
9Qwen 3 VL 8B9#917
10VisualRFT-7B6.8

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

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