Benchmark Almanac

BEAR

Llama-2-13b-hf 66.9% tops its leaderboard Mistral 7B v0.1 65.4% leads the models we track
Not in model cards

Models choose which entity correctly completes a factual relation. The evaluation compares candidate completions across model architectures.

Made by
Jacek Wiland and colleagues
Released
June 2024
Kind
Trivia and open questions
Level
School
Scored
Checked automatically
Answers
Multiple choice
Tasks
All public
Measures
Capability
Language
English
In and out
Text → Text
Interaction
Single turn
Refresh
One fixed set
Run by
Anyone
Strongest model tested
Gemma 7B #211 on the capability index
Last model added
February 2024
Scores refreshed
November 20th, 2026
21
Entries in its paper, 4 models we track

Llama-2-13b-hf tops the leaderboard with 66.9% Mistral 7B v0.1 leads with 65.4% among models released by February 2024

Best score per entry from the benchmark's paper, read on November 20th, 2026. A paper’s table doesn’t add newer models. 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. The best model we track is Mistral 7B v0.1, at 65.4%. This view shows only models we track, each at its best.

The newest model with a score here came out in February 2024, and nobody has tested a newer one. Among the 20 highest-ranked models on the capability index, these have no score yet:

  1. 1 Llama-2-13b-hf 66.9%
  2. 2 Mistral 7B v0.1 65.4%
  3. 3 Gemma 7B 63.7%
  4. 4 Llama-2-7b-hf 62.4%
  5. 5 Gemma 2B 51.5%
  6. 6 opt-30b 47.9%
  7. 7 OPT-1.3B 45.4%
  8. 8 opt-6.7b 43.8%
  9. 9 opt-2.7b 37.3%
  10. 10 gpt2-xl 26.2%
  11. 11 gpt2-large 22.2%
  12. 12 roberta-large 21.5%
  13. 13 bert-large-cased 19.9%
  14. 14 opt-350m 19.6%
  15. 15 gpt2-medium 19%
  16. 16 bert-base-cased 18.4%
  17. 17 opt-125m 16.4%
  18. 18 roberta-base 16.4%
  19. 19 xlm-roberta-large 14.3%
  20. 20 gpt2 13.5%
  21. 21 xlm-roberta-base 11.4%

Source: The BEAR paper

#EntryScoreSettingDateSource
1 Llama-2-13b-hf 66.9% Official leaderboard
2 Mistral 7B v0.1 65.4% Sep 2023 Official leaderboard
3 Gemma 7B 63.7% Feb 2024 Official leaderboard
4 Llama-2-7b-hf 62.4% Official leaderboard
5 Gemma 2B 51.5% Feb 2024 Official leaderboard
6 opt-30b 47.9% Official leaderboard
7 OPT-1.3B 45.4% May 2022 Official leaderboard
8 opt-6.7b 43.8% Official leaderboard
9 opt-2.7b 37.3% Official leaderboard
10 gpt2-xl 26.2% Official leaderboard
11 gpt2-large 22.2% Official leaderboard
12 roberta-large 21.5% Official leaderboard
13 bert-large-cased 19.9% Official leaderboard
14 opt-350m 19.6% Official leaderboard
15 gpt2-medium 19% Official leaderboard
16 bert-base-cased 18.4% Official leaderboard
17 opt-125m 16.4% Official leaderboard
18 roberta-base 16.4% Official leaderboard
19 xlm-roberta-large 14.3% Official leaderboard
20 gpt2 13.5% Official leaderboard
21 xlm-roberta-base 11.4% Official leaderboard
#ModelScoreSettingDateSource
1 Mistral 7B v0.1 65.4% Sep 2023 Official leaderboard
2 Gemma 7B 63.7% Feb 2024 Official leaderboard
3 Gemma 2B 51.5% Feb 2024 Official leaderboard
4 OPT-1.3B 45.4% May 2022 Official leaderboard

Number one over time

Lab claims count from the day the card came out, independent runs from the model’s release.

Mistral AI
Benchmark paper Not tested since Jun 2024

Source: The BEAR paper

Scores over time

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

Plot with other benchmarks
In the benchmark’s paper Best so far
0% 20% 40% 60% 80% 100% 2022 2023 2024 Near the ceiling (85%) Mistral 7B v0.1 OPT-1.3B 0% 20% 40% 60% 80% 100% 2022 2023 2024 Near the ceiling (85%) Mistral 7B v0.1 OPT-1.3B

Source: Official BEAR leaderboard

No lab model card reports it yet

We read the headline results of 540 launch posts from 57 labs, and none of them list this benchmark.

More in Knowledge

All 332 in Knowledge
  • MMLU-Pro In use Harder 10-option version of MMLU with reasoning-focused questions across 14 subjects. Intern-S2-397B 89.8%
  • MMLU In use 57-subject multiple-choice exam questions from elementary to professional level. Scored by accuracy. Seed1.8 92.3%
  • 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%
  • HLE-Diamond In use Models answer 1,000 cleaned expert questions from Humanity's Last Exam, 500 on reasoning and 500 on knowledge, closed book. Graded on exact answers. Claude Fable 5.2 77.9%

Data audit

9 of 10 checks pass, 1 warning. Look first at the Newest score check. The newest came 2 years ago, in February 2024, for Gemma 2B. Every benchmark

Catalog record 2 pass
  • Pass Task access All tasks are public.
  • Pass Confidence Checked against a primary source.
Scores 1 warning 3 pass 1 note
  • Pass Independent runs Scores from the benchmark’s paper (21 entries).
  • Pass Models scored 4 models we track, plus 17 other entries on its leaderboard.
  • Warning Newest score The newest came 2 years ago, in February 2024, for Gemma 2B.
  • Note Frontier models None of the top 10 has a score. The strongest tested is Gemma 7B (#211).
  • Pass Headroom The best is 65.4%, under 85%, so it still separates models.
Leaderboard import 3 pass
  • Pass Importer We read the board on aclanthology.org: 5 rows for models we track and 17 other entries at the last read.
  • Pass Last read November 20th, 2026.
  • Pass Read worked The last read worked.
Consistency 1 pass 1 note
  • Note Lab claims No model has both a lab number and an independent run.
  • Pass Known issues No published audit found problems.

How we check