High Stakes factuality prompts
Models answer difficult medical, legal and financial prompts. A grader with web access counts responses and claims with factual errors.
In-house OpenAI evaluation. Error rates are not production prevalence.
- Made by
- OpenAI
- Released
- August 2026
- Kind
- Hallucination rates
- Level
- Expert
- Scored
- Graded by a model
- Answers
- Open-ended
- Tasks
- Kept private
- Strongest model tested
- GPT-6 Sol #12 on the capability index
- Last in a system card
- October 2026
- Last model added
- October 2026
GPT-6 Sol leads with 4.10% among models released by September 2026
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.
The newest model with a score here came out in September 2026, and nobody has reported a newer one. Among the 20 highest-ranked models on the capability index, these have no score yet:
- Claude Fable 5.2
-
Gemini 4 Carbon - GPT-6.1 Sol
- GPT-6.1 Terra
- and 15 more
- 1 GPT-6 Sol 4.10%
- 2 GPT-6 Luna 4.50%
- 3 GPT-5.5 Instant 19.9%
Source: Labs’ model cards
| # | Model | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | GPT-6 Sol | 4.10% | Oct 2026 | Lab model card | |
| 2 | GPT-6 Luna | 4.50% | Oct 2026 | Lab model card | |
| 3 | GPT-5.5 Instant | 19.9% | May 2026 | Lab model card |
2 models have held number one in labs’ own numbers since its release in August 2026
Lab claims count from the day the card came out, independent runs from the model’s release.
Source: Labs’ model cards
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.
6 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 |
|---|---|---|---|---|
| Oct 29, 2026 | GPT-6 Luna system card addendum | 7.8 | High Stakes, responses with factual errors (claims with errors 1.6%), printed chart label (Figure 3) | Hallucination |
| Oct 7, 2026 | GPT-6 Sol system card addendum | 4.1 | High Stakes, responses with factual errors (claims with errors 0.9%), printed chart label (Figure 3) | Hallucination |
| Oct 7, 2026 | GPT-6 Luna system card addendum | 4.5 | High Stakes, responses with factual errors (claims with errors 1.1%), printed chart label (Figure 3) | Hallucination |
| Aug 6, 2026 | GPT-5.6 Sol system card addendum | 0.18 | High Stakes, relative hallucination rate vs GPT-5.3 Instant, % responses with factual error, 0.18x, printed chart label (Figure 2, 6) | Hallucination |
| Aug 6, 2026 | GPT-5.6 Luna system card addendum | 0.21 | High Stakes, relative hallucination rate vs GPT-5.3 Instant, % responses with factual error, 0.21x, printed chart label (Figure 2, 6) | Hallucination |
| May 4, 2026 | GPT-5.5 Instant system card | 19.9 | High Stakes, responses with factual error (claims with factual error 4.8%), printed chart label (Figure 2) | 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