Benchmark
TruthfulQA
Factuality benchmarks measure how often models get facts right. Scores are percentages.
All factuality benchmarksScores from LLM Stats data.
- Models measured
- 13
- Scores recorded
- 13
- Sources
- 1
- IQ weight
- 1.54
Score distribution
- 60.4%
- Mean
- 57.8%
- Median
- 10.5%
- Std dev
- 50.6%
- 10th percentile
- 77.5%
- 90th percentile
- 50.3%
- Min
This benchmark has a weight of 1.54 for chat models, measured over 13 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 | Mai Thinking 1· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 88% |
| 2 | phi-3.5-moe-instruct· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 77.5% |
| 3 | Phi 4 Mini· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 66.4% |
| 4 | Llama 3.1 Nemotron 70B Instruct· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 58.6% |
| 5 | Qwen 2.5 14B Instruct· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 58.4% |
| 6 | Jamba 1.5 Large· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 58.3% |
| 7 | Qwen2.5 32B Instruct· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 57.8% |
| 8 | Command R Plus· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 56.3% |
| 9 | Qwen2 72B Instruct· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 54.8% |
| 10 | Qwen 2.5 Coder 32B Instruct· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 54.2% |
| 11 | Jamba 1.5 Mini· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 54.1% |
| 12 | Qwen2.5 Coder 7B Instruct· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 50.6% |
| 13 | Mistral Nemo· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 50.3% |
Data credits: Data by LLM Stats