ARC-c
ARC Challenge science questions (Gemma 3n PT card)
knowledge and reasoning
Score in %
Definition
- Version
- not in the notes
- Category
- knowledge and reasoning
- Maintainer
- not in the notes
- Status
- not in the notes
- Home page
- not in the notes
Sources
- Gemma 3n model card | Google AI for Developers ai.google.dev https://ai.google.dev/gemma/docs/gemma-3n/model_card
Scores at a Glance
Score in %
Higher is better for this unit.
1
Jamba 1.5 MiniAI21 Labs
Command R+ (Apr '24)Cohere, Cohere LabsOpen LLM Leaderboard figures as quoted on the model card
DBRXDatabricksDBRX Instruct; 25-shot
Gemma 3n E4BGoogle DeepMindPT (pretrained checkpoint)Gemma 3n E2BGoogle DeepMindPT (pretrained checkpoint)
Jamba 1.5 MiniAI21 Labs85.7 %
vendor-reported2Command R+ (Apr '24)Cohere, Cohere LabsOpen LLM Leaderboard figures as quoted on the model card70.99 %
vendor-reported3DBRXDatabricksDBRX Instruct; 25-shot68.9 %
vendor-reported4Gemma 3n E4BGoogle DeepMindPT (pretrained checkpoint)61.6 %
vendor-reported5Gemma 3n E2BGoogle DeepMindPT (pretrained checkpoint)51.7 %
vendor-reported0255075100
Where this comes from
Each bar is one model, with its most independent recorded score. Hatched bars are vendor-reported; the word beside each bar says who ran it. Settings and dates differ between rows; the Scores table below has every result.
Scores
| Model | Score | Who ran it | Setting | View or population | Date | Source |
|---|---|---|---|---|---|---|
| Jamba 1.5 Mini | 85.7 % | vendor-reported | – | – | Aug 2024 | huggingface.co |
| Command R+ (Apr '24) | 70.99 % | vendor-reported | Open LLM Leaderboard figures as quoted on the model card | – | Apr 2024 | huggingface.co |
| DBRX | 68.9 % | vendor-reported | DBRX Instruct; 25-shot | – | 27 Mar 2024 | databricks.com |
| Gemma 3n E4B | 61.6 % | vendor-reported | PT (pretrained checkpoint) | – | – | ai.google.dev |
| Gemma 3n E2B | 51.7 % | vendor-reported | PT (pretrained checkpoint) | – | – | ai.google.dev |
Where this comes from
Scores from different settings, harnesses or dates are not directly comparable; the row says which. Sorted by score, highest first, as a reading aid, not a ranking.
To cite this page: Project AURION, AI Use-Cases Research & Innovation, "ARC-C", https://aurion.semai.id/benchmarks/arc-c/, records of October 2026.