Nejumi 4 - BFCL - Irrelevance Detection: leaderboard
Metric: Accuracy (%). Source: nejumi.ai. 137 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Llama 4 Scout | 100 |
| 2 | Gemini 2.5 Flash | 96.67 |
| 3 | Phi-4 | 96 |
| 4 | gemma-4-E4B-it | 95 |
| 5 | Gemma 4 E2B (IT) | 95 |
| 6 | DeepSeek V3.2 (Thinking) | 95 |
| 7 | GPT-4.1 | 93.33 |
| 8 | GPT-5 (High) | 93.33 |
| 9 | Grok 3 Mini | 93.33 |
| 10 | GPT-5 Nano (High) | 93.33 |
| 11 | GPT-5.1 (High) | 93.33 |
| 12 | GPT-5.4 (xHigh) | 93.33 |
| 13 | llm-jp-4-32B-a3B (Thinking) | 93.33 |
| 14 | GPT-OSS-Swallow-120B-RL-v0.1 (Thinking) | 93.33 |
| 15 | GPT-4.1 Mini | 91.67 |
Interactive version: theaggregate.ai/benchmark?slug=nejumi-4-bfcl-irrelevance-detection · How It Works · Data refreshed daily, snapshot 2026-09-19.