BenchmarkK2 Horizon 3.7B
- Institute of Foundation Models
- Rilascio · 3 settembre 2026
- Pesi aperti
- Ragionamento
Intelligenza
Intelligence Index
#120 su 266- 17,718
Qwen3.5 122B A10B (Non-reasoning) - 17,518
Muse Glimmer (High) - 16,917
Doubao Seed Code - 16,717
Gemma 4 26B A4B (Reasoning) - 16,617
Ring-2.6-1T - 15,616K2 Horizon 3.7B
- 15,616
Qwen3.5 122B A10B (Reasoning) - 15,215
Qwen3.6 35B A3B (Non-reasoning) - 14,715
Gemma 4 31B (Reasoning) - 14,314
ERNIE 5.0 Thinking Preview
Dati di Artificial Analysis- AA-Omniscience Accuracy10,8%#215
- Humanity's Last Exam13,9%#122
- Terminal-Bench 4.00%#96
- AA-LCR62,3%#126
- SciCode22%#133
- CritPt0%#143
- GPQA Diamond69,2%#121
- τ²-Bench Banking19,6%#48
K2 Horizon 3.7BMediana
Dati di Artificial AnalysisModelli vicini