Nejumi 4 - BFCL - Relevance Detection: leaderboard
Metric: Accuracy (%). Source: nejumi.ai. 137 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Llama 4 Maverick Instruct FP8 | 100 |
| 2 | Grok 4.5 (High) | 100 |
| 3 | Grok 4.20 0309 (Reasoning) | 100 |
| 4 | qwen2.5-bakeneko-32B-instruct-v2 | 100 |
| 5 | ELYZA-Shortcut-1.0-Qwen-32B | 100 |
| 6 | Gemma 3 12B (IT) | 94.44 |
| 7 | Qwen 3 Next 80B A3B Instruct | 94.44 |
| 8 | Qwen 3 235B A22B 2507 (Thinking) | 94.44 |
| 9 | Llama 3.1 405B Instruct FP8 | 94.44 |
| 10 | Llama-3.3-Swallow-70B-Instruct-v0.4 | 94.44 |
| 11 | GPT-4o Mini (2024-07-18) | 88.89 |
| 12 | Grok 4 | 88.89 |
| 13 | Qwen 3 30B A3B 2507 (Thinking) | 88.89 |
| 14 | DeepSeek V3.2 (Thinking) | 88.89 |
| 15 | DeepSeek V4 Pro (Max) | 88.89 |
Interactive version: theaggregate.ai/benchmark?slug=nejumi-4-bfcl-relevance-detection · How It Works · Data refreshed daily, snapshot 2026-09-19.