List your company

Question Answering on ARC-Challenge (Accuracy, TFLOPs)

94.37Accuracydense alignment
85.488487.794290.192.4058Jun 11, 2026

Evaluation results

MethodLinks
dense alignmentPair=8B → 14B, Regime=context-aware2026.0694.3721.79
dense alignmentPair=4B → 14B, Regime=context-aware2026.0694.221.54
dense alignmentPair=14B → 8B, Regime=context-aware2026.0693.7715.64
dense alignmentPair=4B → 8B, Regime=context-aware2026.0693.6912.56
dense alignmentPair=8B → 4B, Regime=context-aware2026.06937.95
Receiver-onlyPair=4B → 14B, Regime=context-aware2026.0692.631.16
Receiver-onlyPair=8B → 14B, Regime=context-aware2026.0692.631.16
C2CPair=4B → 14B, Regime=context-aware2026.0692.436.64
C2CPair=8B → 14B, Regime=context-aware2026.0692.437.42
T2TPair=14B → 8B, Regime=context-aware2026.0692.3555.5
T2TPair=4B → 14B, Regime=context-aware2026.069256.24
dense alignmentPair=14B → 4B, Regime=context-aware2026.0691.8910.18
T2TPair=4B → 8B, Regime=context-aware2026.0691.7437.73
T2TPair=8B → 14B, Regime=context-aware2026.0691.7467.08
Receiver-onlyPair=4B → 8B, Regime=context-aware2026.069119.85
Receiver-onlyPair=14B → 8B, Regime=context-aware2026.069119.85
T2TPair=8B → 4B, Regime=context-aware2026.0690.9633.91
T2TPair=14B → 4B, Regime=context-aware2026.0689.8343.48
C2CPair=14B → 8B, Regime=context-aware2026.0689.397.83
Receiver-onlyPair=8B → 4B, Regime=context-aware2026.0689.29.18
Receiver-onlyPair=14B → 4B, Regime=context-aware2026.0689.29.18
C2CPair=8B → 4B, Regime=context-aware2026.0686.783.43
C2CPair=4B → 8B, Regime=context-aware2026.0686.094.5
C2CPair=14B → 4B, Regime=context-aware2026.0685.835.01