Multi-factual (Anthropic)
Models answer questions that each combine several closed-ended sub-questions on one topic. Answers are graded correct, incorrect or unsure against references.
In-house Anthropic factuality evaluation. Questions were built from article quotes and checked by hand. Score is the share answered correctly.
- Made by
- Anthropic
- Released
- March 2024
- Kind
- Long-form factuality
- Level
- Everyday
- Scored
- Graded by a model
- Answers
- Open-ended
- Tasks
- Kept private
- Strongest model tested
- Claude 3 Opus #172 on the capability index
- Last in a system card
- March 2024
- Last model added
- March 2024
Claude 3 Opus leads with 62.8%
Best score per model as labs report them in their own cards. Labs pick their own settings, so these compare less cleanly than independent runs.
- 1 Claude 3 Opus 62.8%
Source: Labs’ model cards
| # | Model | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | Claude 3 Opus | 62.8% | Mar 2024 | Lab model card |
Scores over time
All the scores we have, one dot each, so a model tested at several settings appears more than once.
Source: Labs’ model cards
Labs report it only in system cards
We read the headline results of 522 launch posts from 52 labs, and none of them list this benchmark. Some system cards do, below.
3 system cards report it, from 1 lab
Besides launch posts, we read the system cards, model cards and technical reports labs publish with a release. They report far more, including safety and frontier-risk tests.
| Date | Document | Score | As the document writes it | Section |
|---|---|---|---|---|
| Mar 4, 2024 | Claude 3 Sonnet model card | Multi-factual, correct, incorrect and unsure shares, chart without printed values (Figure 11) | Hallucination | |
| Mar 4, 2024 | Claude 3 Opus model card | 62.8 | Multi-factual, % correct ("over 62.8%"), stated in text. Incorrect and unsure shares in a chart without printed values (5.7, Figure 11) | Hallucination |
| Mar 4, 2024 | Claude 3 Haiku model card | Multi-factual, correct, incorrect and unsure shares, chart without printed values (Figure 11) | Hallucination |
More in Factuality
- SimpleVQA In use Models answer 2,025 short factual questions about images across 9 tasks and 9 topics. An LLM judge grades answers against fixed references.
Qwen3.7-Plus 81.7% - HallusionBench In use Models answer 1,129 yes or no questions about 346 images, built in pairs to expose language hallucination and visual illusion. Scored by accuracy.
Keye-VL-2.0-30B-A3B 73.5% - SimpleQA Verified In use 1,000 human-verified, de-duplicated short-answer factual questions measuring parametric knowledge.
Gemini 3.1 Pro 77.5% - Kimi Hallu Bench In use Models answer hard prompts drawn from practical work tasks. Fact-checking agents score the share of claims that are hallucinated. GPT-6 Astra 4.20%
- Internal adversarially-designed factuality evaluations In use Models answer adversarial questions in nine categories such as history, math and code. Scored on agreement with human ideal responses. GPT-6 Luna 7.60%
- AA-Omniscience v1.1 In use Models answer 3,000 knowledge questions in six domains without tools and may abstain. The index adds correct answers and subtracts wrong ones.
Inkling 2 Small 16.1