Benchmark Almanac

Hallucinations on production-representative prompts

GPT-5.2 3.10% Only in system cards

Models answer prompts representative of ChatGPT production traffic. A grader with web access scores incorrect claims and responses with major errors.

In-house OpenAI evaluation. Cards report results with and without browsing.

Made by OpenAI Released Aug 2025 Hallucination rates Tasks kept private Last in a system card Dec 2025 Last model added Dec 2025
Level
Everyday Most adults, without training
Scored
Graded by a model A language model scores each response, usually against a rubric.
Answers
Open-ended Free-form text, documents, designs or media judged as a whole.
2
Models scored in lab cards
2
System cards report it, from 1 lab

GPT-5.2 leads with 3.10%

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. 1 GPT-5.2 3.10%
  2. 2 GPT-5 4.50%
#ModelScoreSettingDateSource
1 GPT-5.2 3.10% Dec 2025 Model card
2 GPT-5 4.50% Aug 2025 Model card

Scores over time

All the scores we have, one dot each, so a model tested at several settings appears more than once.

Reported by the lab in its model card Best so far
0% 20% 40% 60% 80% 100% 2026 GPT-5.2 GPT-5 0% 20% 40% 60% 80% 100% 2026 GPT-5.2

Labs report it only in system cards

We read the headline results of 479 launch posts from 51 labs, and none of them list this benchmark. Some system cards do, below.

2 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.

DateDocumentScoreAs the document writes itSection
11 Dec 2025 GPT-5.2 system card addendum 3.1 Average Hallucination Rate (Browsing Disabled), % incorrect claims, gpt-5.2-thinking, printed chart label (Figure 2). Responses with 1+ major incorrect claims 10.9%. Browsing enabled 0.8% and 5.8% (Figure 1) Hallucination
7 Aug 2025 GPT-5 system card 4.5 Factuality on ChatGPT Production Traffic (Browsing Enabled), % incorrect claims (responses with 1+ major incorrect claims 4.8%) (Figure 1) Hallucination

More in Factuality

All 70 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%
  • GaRAGe In use Models write long-form RAG answers to 2,366 questions with human-annotated grounding passages. Scoring checks whether answers use the relevant passages. Nemotron 3.5 Bolt E4B 40.7%
  • 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. Kimi K3.1 Pro 8.30%
  • RAGTruth In use A word-level annotated corpus of hallucinations in retrieval-augmented generation answers. It is used to train and evaluate hallucination detection and RAG faithfulness. Nemotron 3.5 Bolt E4B 77.9%