K2 Horizon 3.7B: benchmark results
Provider: Institute of Foundation Models. Released 2026-09-03. Access: Open.
Aggregate Rating 1587 ± 23, rank #754 of 2171 rated models, from 41 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| AA Humanity's Last Exam | 13.86 | Accuracy (%) | 58 |
| AA Long Context Reasoning | 62.33 | Accuracy (%) | 51 |
| Tau3 Banking | 19.59 | Success Rate (%) | 50.7 |
| AA GPQA Diamond | 69.19 | Accuracy (%) | 46.8 |
| LLM Stats Score | 21.37 | LLM Stats Score (conservative rating) | 46.7 |
| AA-Briefcase - PDF - Rubric Pass Rate | 25.86 | Rubric Pass Rate (%) | 41.4 |
| ZeroEval GPQA Diamond | 68.9 | GPQA Diamond Score | 39.1 |
| AA GDPval | 889.22 | ELO | 36.8 |
| AutomationBench-AA - Support - Raw Objective Completion | 49.02 | Raw Objective Completion (%) | 35.9 |
| AutomationBench-AA | 14.7 | Guardrail-Adjusted Objective Completion (%) | 34.2 |
| AA-Briefcase - XLSX - Rubric Pass Rate | 19.32 | Rubric Pass Rate (%) | 34 |
| AutomationBench-AA - Sales - Raw Objective Completion | 32.39 | Raw Objective Completion (%) | 33.8 |
Interactive version: theaggregate.ai/model?slug=k2-horizon-3-7b · How It Works · Data refreshed daily, snapshot 2026-10-30.