Codesota · Benchmark · ImageNet Zero-ShotHome/Leaderboards/ImageNet Zero-Shot
Unknown

ImageNet Zero-Shot.

Zero-shot classification accuracy on ImageNet without task-specific training

Paper ↗Leaderboard ↓
§ 01 · SOTA history

Year over year.

§ 02 · Leaderboard

Results by metric.

Only 4 models on this benchmark
Help build the community leaderboard — submit your model results.
Found a wrong score or missing run?
Use row edits to send a sourced correction into moderation.
Add / edit result ↗Report issue ↗

Top 1

Top 1 is the reported evaluation metric for ImageNet Zero-Shot. Codesota tracks published model scores on this metric so readers can compare state-of-the-art results across sources and model families.

Higher is better

Trust tiers for Top 1verifiedpapervendorcommunityunverified

Muted rows were not state of the art when published — an earlier or same-year result already scored better.

RankModelTrustScoreYearLinksFix
01EVA-CLIP-18B
EVA-CLIP-18B (18 billion parameters). Zero-shot ImageNet top-1. Table 1. BAAI/BAAIvision, 2024.
verified83.82024Source ↗Looks wrong?
02SigLIP-SO400M
SigLIP SO400M/14@384 model. Zero-shot ImageNet top-1 accuracy. Google Brain, ICCV 2023. Table 4.
verified83.22023Source ↗Looks wrong?
03OpenCLIP ViT-G/14
OpenCLIP ViT-G/14 trained on LAION-2B. Zero-shot ImageNet top-1. Table 2.
verified80.12022Source ↗Looks wrong?
04CLIP ViT-L/14
CLIP ViT-L/14. Zero-shot top-1 accuracy on ImageNet. OpenAI, ICML 2021. Table 5.
verified75.52021Source ↗Looks wrong?
§ 04 · Submit a result

Add to the leaderboard.

← Back to Leaderboards
ImageNet Zero-Shot Leaderboard | CodeSOTA | CodeSOTA