Syslog Severity Classification - Few-Shot: leaderboard
Small and small reasoning language models label real Linux journalctl log entries with their syslog severity level (0-7) with five labelled example logs from the training split in the prompt; malformed or verbose answers count as wrong. Accuracy in percent on 9,355 held-out logs, higher is better; Masri et al. (George Mason University), 2026.
Metric: Accuracy on 9,355 held-out journalctl logs, exact syslog severity level (%). Source: arxiv.org. Saturation forecast: Estimated already saturated. 9 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Qwen 3 4B | 56.01 | #2297 |
| 2 | Qwen 3 1.7B | 43.3 | #5551 |
| 3 | Gemma 3 4B | 41.06 | #5104 |
| 4 | Qwen 3 0.6B | 28.92 | #8392 |
| 5 | Gemma 3 1B | 20.25 | #9004 |
| 6 | DeepSeek R1 Distill Qwen 1.5B | 17.63 | #9257 |
| 7 | Phi-4-mini (Reasoning) | 0 |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=syslog-severity-classification-few-shot · How It Works · Data refreshed daily, snapshot 2026-11-12.