Benchmark
DROP
Scores are percentages. Higher is better.
- Models and systems measured
- 24
- Scores recorded
- 24
- Sources
- 2
- IQ weight
- 2.04
Score distribution
- 78.7%
- Mean
- 80.9%
- Median
- 12.2%
- Std dev
- 70.2%
- 10th percentile
- 87.1%
- 90th percentile
- 28.6%
- Min
This benchmark has a weight of 2.04 for chat models, measured over 22 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.
Leaderboard
| # | Model or system | Score |
|---|---|---|
| 1 | DeepSeek V3· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 91.6% |
| 2 | Claude 3.5 Sonnet· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 87.1% |
| 3 | MiMo V2.5 Pro· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 86.3% |
| 4 | GPT-4 Turbo· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 86% |
| 5 | Nova Pro· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 85.4% |
| 6 | Llama 3.1 405B Instruct· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 84.8% |
| 7 | GPT 4o· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 83.4% |
| 8 | Claude 3 Opus· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 83.1% |
| 9 | Claude Haiku 3.5· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 83.1% |
| 10 | GPT 4· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 80.9% |
| 11 | Nova Lite· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 80.2% |
| 12 | GPT-4o mini· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 79.7% |
| 13 | Llama 3.1 70B Instruct· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 79.6% |
| 14 | Nova Micro· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 79.3% |
| 15 | LongCat Flash Chat· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 79.1% |
| 16 | Claude 3 Sonnet· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 78.9% |
| 17 | Claude 3 Haiku· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 78.4% |
| 18 | Phi 4· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 75.5% |
| 19 | Gemini 1.5 Pro· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 74.9% |
| 20 | GPT-3.5 Turbo· LLM Stats data | 70.2% |
| 21 | Llama 3.1 8B Instruct· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 59.5% |
| 22 | ERNIE 4.5· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 28.6% |
Not ranked
The source marks these results as descriptive, not capability scores, so we list them by name instead of ranking them.
| Model or system | Score |
|---|---|
| DeepSeek V4.1 Flash BaseF1 · 1-shot· Reported by DeepSeek· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 87.9% |
| MiMo V2 Flash Baseofficial Base table · 3-shot· Reported by Xiaomi MiMo· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 84.7% |
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.