BLiMP
Language models must assign higher probability to the grammatical sentence in each of 67,000 minimal pairs covering English syntax, morphology and semantics.
67 sub-datasets of 1,000 pairs. Core evaluation of the BabyLM challenge.
- Released
- December 2019
- Kind
- Commonsense
- Level
- Everyday
- Scored
- Checked automatically
- Answers
- Multiple choice
- Tasks
- All public
In the authors’ words
We introduce The Benchmark of Linguistic Minimal Pairs (shortened to BLiMP), a challenge set for evaluating what language models (LMs) know about major grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each containing 1000 minimal pairs isolating specific contrasts in syntax, morphology, or semantics. The data is automatically generated according to expert-crafted grammars, and aggregate human agreement with the labels is 96.4%. We use it to evaluate n-gram, LSTM, and Transformer (GPT-2 and Transformer-XL) LMs. We find that state-of-the-art models identify morphological contrasts reliably, but they struggle with semantic restrictions on the distribution of quantifiers and negative polarity items and subtle syntactic phenomena such as extraction islands.
BLiMP gpt2 family tops the leaderboard with 83%
Best score per entry from the benchmark's own leaderboard, read on November 12th, 2026. Where a model ran with several settings or harnesses, we keep its best. Names in gray aren't in our model list: fine-tunes, agent systems, products and older models. They don't count toward leaders.
- 1 BLiMP gpt2 family Equal-weight paradigm mean 83%
- 2 BLiMP lstm family Equal-weight paradigm mean 69.8%
- 3 BLiMP txl family Equal-weight paradigm mean 69.6%
- 4 BLiMP ngram family Equal-weight paradigm mean 61.2%
Source: Official BLiMP leaderboard
| # | Entry | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | BLiMP gpt2 family | 83% | Equal-weight paradigm mean | Official leaderboard | |
| 2 | BLiMP lstm family | 69.8% | Equal-weight paradigm mean | Official leaderboard | |
| 3 | BLiMP txl family | 69.6% | Equal-weight paradigm mean | Official leaderboard | |
| 4 | BLiMP ngram family | 61.2% | Equal-weight paradigm mean | Official leaderboard |
No lab model card reports it yet
We read the headline results of 522 launch posts from 52 labs, and none of them list this benchmark.
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%