Skip to content

Benchmark

TruthfulQA

Factuality benchmarks measure how often models get facts right. Scores are percentages.

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
50.3%88%
The band covers the middle 80% of scores. The line marks the median.

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

#ModelScore
1Mai Thinking 1· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim88%
2phi-3.5-moe-instruct· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim77.5%
3Phi 4 Mini· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim66.4%
4Llama 3.1 Nemotron 70B Instruct· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim58.6%
5Qwen 2.5 14B Instruct· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim58.4%
6Jamba 1.5 Large· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim58.3%
7Qwen2.5 32B Instruct· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim57.8%
8Command R Plus· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim56.3%
9Qwen2 72B Instruct· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim54.8%
10Qwen 2.5 Coder 32B Instruct· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim54.2%
11Jamba 1.5 Mini· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim54.1%
12Qwen2.5 Coder 7B Instruct· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim50.6%
13Mistral Nemo· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim50.3%

Data credits: Data by LLM Stats