Question Answering on ARC Easy (Accuracy and Modality Gap)
95.9AccuracyTextPro-SLM-7B 5:1
Evaluation results
| Method | Links | ||
|---|---|---|---|
| TextPro-SLM-7B 5:1System Category=Modality Gap-optimized Systems, Sampling Ratio=5:12026.05 | 95.9 | 0.2 | |
| TextPro-SLM-7BSystem Category=Modality Gap-optimized Systems2026.05 | 95.8 | 0.3 | |
| Kimi-Audio-7BSystem Category=End-to-end Systems, Backbone LLM=Kimi-Audio-7B2026.05 | 95.2 | 0.9 | |
| WhisperPro + Qwen2.5-7BSystem Category=Ablation Studies, Backbone LLM=Qwen2.5-7B2026.05 | 95 | 1.1 | |
| Qwen2.5-Omni-7BSystem Category=End-to-end Systems, Backbone LLM=Qwen2.5-Omni-7B2026.05 | 94.9 | 1.2 | |
| TextPro-SLM-3BSystem Category=Modality Gap-optimized Systems2026.05 | 92.9 | 1.4 | |
| ASR + Qwen2.5-3BSystem Category=Cascaded Toplines, ASR=Whisper-v3-Large, Backbone LLM=Qwen2.5-3B2026.05 | 90.5 | 3.7 | |
| ASR + Qwen2.5-7BSystem Category=Cascaded Toplines, ASR=Whisper-v3-Large, Backbone LLM=Qwen2.5-7B2026.05 | 89.9 | 6.2 | |
| TextPro-SLM-7B w/o KDSystem Category=Ablation Studies, Ablation Setting=w/o KD2026.05 | 89.5 | 6.6 | |
| GLM-4-Voice-9BSystem Category=End-to-end Systems, Backbone LLM=GLM-4-Voice-9B2026.05 | 73.2 | 24.6 | |
| Qwen2-Audio-7BSystem Category=End-to-end Systems, Backbone LLM=Qwen2-Audio-7B2026.05 | 43.5 | 28.5 | |
| DiVA-Llama3.1-8BSystem Category=End-to-end Systems, Backbone LLM=Llama3.1-8B2026.05 | 39.7 | 53.6 | |
| Random2026.05 | 25 | — |