Benchmark
IFBench
Instruction following benchmarks measure how closely models follow instructions. Scores are percentages. Higher is better.
- Models measured
- 51
- Scores recorded
- 70
- Sources
- 2
- IQ weight
- 1.40
Score distribution
- 71.6%
- Mean
- 74.2%
- Median
- 11.2%
- Std dev
- 59.2%
- 10th percentile
- 81.3%
- 90th percentile
- 21%
- Min
This benchmark has a weight of 1.40 for chat models, measured over 51 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.
Leaderboard
| # | Model | Score |
|---|---|---|
| 1 | Qwen3.8 Max· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 82.8% |
| 2 | Qwen3.8 MaxQwen3.8 官方模型卡对照表· Reported by Qwen· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 82.8% |
| 3 | Inkling Small· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 82.2% |
| 4 | Nemotron 3 Ultra· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 81.7% |
| 5 | Nemotron 3 UltraNeMo Evaluator SDK · prompt loose· Reported by NVIDIA· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 81.7% |
| 6 | Qwen3.8 Omni FlashQwen3.8 Omni Flash 官方发布表· Reported by Qwen· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 81.5% |
| 7 | Qwen 3.8 Flash Next· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 81.3% |
| 8 | Qwen 3.8 Flash NextQwen3.8-Flash-Next release/model card · published setting· Reported by Qwen· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 81.3% |
| 9 | Qwen3.8 Flash· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 81.3% |
| 10 | Qwen3.8 FlashQwen3.8 Omni Flash 官方发布表· Reported by Qwen· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 81.3% |
| 11 | Inkling FP4· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 79.8% |
| 12 | Beam 1.0 501B A23B· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 79.6% |
| 13 | Qwen 3.8 27B· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 79.5% |
| 14 | Qwen 3.8 27BQwen3.8 Omni Flash 官方发布表· Reported by Qwen· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 79.5% |
| 15 | Qwen 3.8 27BQwen3.8-27B model card · published setting· Reported by Qwen· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 79.5% |
| 16 | Granite 4.2 8B· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 79.3% |
| 17 | DeepSeek V4 Flash 0731Qwen3.8 Omni Flash 官方发布表· Reported by Qwen· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 79.2% |
| 18 | Qwen3.7 Max· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 79.1% |
| 19 | Qwen3.7 MaxQwen3.7-Max release · recommended reasoning-effort=xhigh system prompt· Reported by Qwen· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 79.1% |
| 20 | Qwen3.7 MaxQwen3.8 官方模型卡对照表· Reported by Qwen· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 79.1% |
| 21 | Qwen3.7 Plus· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 79.1% |
| 22 | Qwen3.7 PlusQwen3.7-Plus release · recommended reasoning-effort=xhigh system prompt· Reported by Qwen· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 79.1% |
| 23 | Qwen3.7 PlusQwen3.8 Omni Flash 官方发布表· Reported by Qwen· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 79.1% |
| 24 | Granite 4.2 30B· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 77.2% |
| 25 | Muse Glimmer 30B· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 77% |
| 26 | Muse Glimmer 30Bhigh reasoning· Reported by Meta· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 77% |
| 27 | Mercury 3· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 76.8% |
| 28 | Qwen3.5 27B· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 76.5% |
| 29 | Qwen3.5 397B A17B· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 76.5% |
| 30 | Qwen3.5 397B A17BQwen3.5 release/model card · published setting· Reported by Qwen· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 76.5% |
| 31 | Qwen3.5 122B-A10B· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 76.1% |
| 32 | Gemma 4 31B ITGemma 4 Technical Report · Table 5 · thinking mode unless noted· Reported by Google DeepMind· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 76% |
| 33 | Ling 3.0 Flash· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 74.5% |
| 34 | Granite 4.2 3B· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 74.3% |
| 35 | Qwen3.6 Plus· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 74.2% |
| 36 | Qwen3.6 PlusQwen official release table · thinking mode unless benchmark requires otherwise· Reported by Alibaba Qwen· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 74.2% |
| 37 | Command A Plus 05 2026· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 74% |
| 38 | GPT-5.6 SolQwen3.8 官方模型卡对照表· Reported by Qwen· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 72.7% |
| 39 | Nemotron 3 Super 120B A12Bofficial blog Figure 3 · rounded· Reported by NVIDIA Developer· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 72.6% |
| 40 | Nemotron 3 Super 120B A12BNeMo Evaluator / mostly NeMo Skills · prompt · report Table 8 BF16· Reported by NVIDIA· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 72.6% |
| 41 | Nemotron 3 Super 120B A12B· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 72.6% |
| 42 | Nemotron 3 Super 120B A12BNeMo Evaluator / mostly NeMo Skills · prompt· Reported by NVIDIA· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 72.6% |
| 43 | Gemma 4 26B A4BGemma 4 Technical Report · Table 5 · thinking mode unless noted· Reported by Google DeepMind· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 72% |
| 44 | Nemotron 3.5 Lightning 30B A3B· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 71.9% |
| 45 | Nemotron 3.5 Lightning 30B A3Bloose· Reported by NVIDIA· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 71.9% |
| 46 | Mercury 2· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 71% |
| 47 | Qwen3 Max· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 70.9% |
| 48 | NVIDIA Nemotron 3.5 Super VL 120B· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 70.9% |
| 49 | Nova 2 Lite· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 70.8% |
| 50 | Nova 2 LiteAmazon Nova 2 Lite official launch chart / technical report · prompt-loose accuracy, v1· Reported by Amazon Web Services, Amazon AGI· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 70.8% |
Showing 1–50 of 70 results.
Data credits: Data by LLM Stats · BenchBoard by SYuan03 (github.com/SYuan03/benchboard). MIT covers its code and data organization; no results licence is inferred. Numeric relays retain the original stating-source links.