SQuAD v2
Llama 3.1-405B 81.8% leads since November 2024 Models extract answers from Wikipedia paragraphs and must abstain on over 50,000 adversarial unanswerable questions. Scored by exact match and F1.
Saturated. The test set is hidden, so cards report the dev set.
- Made by
- Stanford NLP
- Released
- June 2018
- Kind
- Reading comprehension
- Level
- Everyday
- Scored
- Checked automatically
- Answers
- Short answer
- Tasks
- Public, plus a private test set
- Strongest model tested
- Llama 3.1-405B #162 on the capability index
- Last in a launch post
- December 2023, dropped before it saturated
- Last model added
- November 2024
In the authors’ words
Extractive reading comprehension systems can often locate the correct answer to a question in a context document, but they also tend to make unreliable guesses on questions for which the correct answer is not stated in the context. Existing datasets either focus exclusively on answerable questions, or use automatically generated unanswerable questions that are easy to identify. To address these weaknesses, we present SQuAD 2.0, the latest version of the Stanford Question Answering Dataset (SQuAD). SQuAD 2.0 combines existing SQuAD data with over 50,000 unanswerable questions written adversarially by crowdworkers to look similar to answerable ones. To do well on SQuAD 2.0, systems must not only answer questions when possible, but also determine when no answer is supported by the paragraph and abstain from answering. SQuAD 2.0 is a challenging natural language understanding task for existing models: a strong neural system that gets 86% F1 on SQuAD 1.1 achieves only 66% F1 on SQuAD 2.0.
2 models share the top score of 81.8% among models released by July 2024
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 July 2024, 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 Astra
- Claude Opus 5.5
- and 16 more
- 1 Llama 3.1-405B 81.8%
- 2 Llama 3.1-70B 81.8%
- 3 Llama 2-70B 80.7%
- 4 PaLM 2-L 80.5%
- 5 Llama 2-34B 77.4%
- 6 Llama 3.1-8B 77%
- 7 Llama 2-13B 72.9%
- 8 Llama 2-7B 67.2%
Source: Labs’ model cards
| # | Model | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | Llama 3.1-405B | 81.8% | Nov 2024 | Lab model card | |
| 2 | Llama 3.1-70B | 81.8% | Nov 2024 | Lab model card | |
| 3 | Llama 2-70B | 80.7% | Jul 2023 | Lab model card | |
| 4 | PaLM 2-L | 80.5% | May 2023 | Lab model card | |
| 5 | Llama 2-34B | 77.4% | Jul 2023 | Lab model card | |
| 6 | Llama 3.1-8B | 77% | Nov 2024 | Lab model card | |
| 7 | Llama 2-13B | 72.9% | Jul 2023 | Lab model card | |
| 8 | Llama 2-7B | 67.2% | Jul 2023 | Lab model card |
3 models have held number one in labs’ own numbers since 2023
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
No launch post has listed it since December 2023
Launch posts from every lab that has listed it, one tick per post. Tall colored ticks include it, and short gray ones leave it out.
7 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 |
|---|---|---|---|---|
| Nov 23, 2024 | Llama 3.1-8B technical report | 77 | SQuAD, pretrained (Table 9). Table 8 lists the reading benchmark as SQuAD V2 | Capabilities |
| Nov 23, 2024 | Llama 3.1-70B technical report | 81.8 | SQuAD, pretrained (Table 9). Table 8 lists the reading benchmark as SQuAD V2 | Capabilities |
| Nov 23, 2024 | Llama 3.1-405B technical report | 81.8 | SQuAD, pretrained (Table 9). Table 8 lists the reading benchmark as SQuAD V2 | Capabilities |
| Jul 19, 2023 | Llama 2-7B technical report | 67.2 | SQuAD (EM), 0-shot, cited as Rajpurkar et al. 2018, pretrained (Table 23) | Capabilities |
| Jul 19, 2023 | Llama 2-70B technical report | 80.7 | SQuAD (EM), 0-shot, cited as Rajpurkar et al. 2018, pretrained (Table 23) | Capabilities |
| Jul 19, 2023 | Llama 2-34B technical report | 77.4 | SQuAD (EM), 0-shot, cited as Rajpurkar et al. 2018, pretrained (Table 23) | Capabilities |
| Jul 19, 2023 | Llama 2-13B technical report | 72.9 | SQuAD (EM), 0-shot, cited as Rajpurkar et al. 2018, pretrained (Table 23) | Capabilities |
Versions
More in Knowledge
- MMLU-Pro In use Harder 10-option version of MMLU with reasoning-focused questions across 14 subjects.
Intern-S2-397B 89.8% - SuperGPQA In use 26,529 graduate-level multiple-choice questions across 285 disciplines. Accuracy.
Qwen 3.6 Max (Preview) 73.9% - WideSearch In use 200 bilingual broad info-seeking tasks: agent collects many atomic facts into a table. Scored by success rate and row/item F1.
Tencent Hy4 preview 83.9% - BBEH micro average In use Models solve harder replacements for each of the 23 BIG-Bench Hard tasks, 4,520 examples in total. The score is the share of all examples answered correctly. Gemma 4 31B IT 74.4%
- KMMLU-Pro In use Models answer multiple-choice questions taken from Korean National Professional Licensure exams. It tests professional knowledge used in Korean industry.
Solar Pro 4 79.2% - HLE-Verified In use Verified and revised version of Humanity's Last Exam (1,811 items) with uncertain items removed.
Seed2.0 Pro 73.6%