K2-V2 (High) — benchmark results
Provider: LLM360. Released 2026-01-27. Access: Open.
Unified ELO 1593 ± 22, rank #510 of 1841 rated models, from 53 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| AA LiveCodeBench | 69.42 | Pass@1 (%) | 78.2 |
| AA AIME 2025 | 78.33 | Accuracy (%) | 73.1 |
| AA IFBench | 60.14 | Accuracy (%) | 71.4 |
| AA MMLU-Pro | 78.59 | Accuracy (%) | 61 |
| AA Humanity's Last Exam | 9.82 | Accuracy (%) | 60.8 |
| AA Omniscience - Software Engineering (SWE) - Kotlin | 20 | Accuracy (%) | 54.3 |
| AA-LCR | 33.3 | Score (self-reported) | 52.4 |
| AA Omniscience - Health | 19.1 | Accuracy (%) | 49.7 |
| AA GPQA Diamond | 68.08 | Accuracy (%) | 49.1 |
| AA Omniscience - Science, Engineering & Mathematics | 25.8 | Accuracy (%) | 48.5 |
| AA Omniscience - Software Engineering (SWE) - JavaScript | 29.09 | Accuracy (%) | 48.4 |
| AA Omniscience - Software Engineering (SWE) - HTML | 30 | Accuracy (%) | 47.3 |
Interactive version: theaggregate.ai/model?slug=k2-v2-high · How It Works · Data refreshed daily, snapshot 2026-07-25.