Benchmark Almanac

BRoverbs

Claude-3.5-sonnet (snapshot unspecified) 97% tops its leaderboard Claude 3.5 Haiku 95% leads the models we track
Not in model cards

Models select the Brazilian proverb that fits a story or the story that illustrates a proverb. Both directions offer five choices.

The release contains 579 questions in each direction. PoETa V2 reports separate model scores for both directions.

Released
September 2025
Kind
Reasoning in many languages
Level
Everyday
Scored
Checked automatically
Answers
Multiple choice
Tasks
All public
Measures
Capability
Language
Portuguese
In and out
Text → Text
Interaction
Single turn
Refresh
One fixed set
Run by
Anyone
Strongest model tested
Claude 3.5 Haiku #175 on the capability index
Last model added
October 2024
Scores refreshed
November 20th, 2026
19
Entries in its paper, 8 models we track

Claude-3.5-sonnet (snapshot unspecified) tops the leaderboard with 97% Claude 3.5 Haiku leads with 95% among models released by October 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 Claude 3.5 Haiku, at 95%. This view shows only models we track, each at its best.

The newest model with a score here came out in October 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 Claude-3.5-sonnet (snapshot unspecified) 97%
  2. 2 GPT-4o (snapshot unspecified) 97%
  3. 3 Sabiá-3 96%
  4. 4 GPT-4o-mini (snapshot unspecified) 95%
  5. 5 Sabiazinho-3 95%
  6. 6 Claude 3.5 Haiku 95%
  7. 7 Qwen2.5-14B 93%
  8. 8 Qwen2.5-7B 83%
  9. 9 Llama 3-8B 61%
  10. 10 Qwen2.5-3B 59%
  11. 11 Qwen2.5-1.5B 49%
  12. 12 Sabiá-7B 28%
  13. 13 Llama 2-7B 23%
  14. 14 LLaMA-7B 22%
  15. 15 Curió 1.1B 20%
  16. 16 TinyLlama 1T 20%
  17. 17 TinyLlama 3T 19%
  18. 18 Tucano 1.1B 19%
  19. 19 Tucano 2.4B 19%

Source: The BRoverbs paper

#EntryScoreSettingDateSource
1 Claude-3.5-sonnet (snapshot unspecified) 97% Official leaderboard
2 GPT-4o (snapshot unspecified) 97% Official leaderboard
3 Sabiá-3 96% Official leaderboard
4 GPT-4o-mini (snapshot unspecified) 95% Official leaderboard
5 Sabiazinho-3 95% Official leaderboard
6 Claude 3.5 Haiku 95% Oct 2024 Official leaderboard
7 Qwen2.5-14B 93% Sep 2024 Official leaderboard
8 Qwen2.5-7B 83% Sep 2024 Official leaderboard
9 Llama 3-8B 61% Apr 2024 Official leaderboard
10 Qwen2.5-3B 59% Sep 2024 Official leaderboard
11 Qwen2.5-1.5B 49% Sep 2024 Official leaderboard
12 Sabiá-7B 7B 28% Official leaderboard
13 Llama 2-7B 23% Jul 2023 Official leaderboard
14 LLaMA-7B 22% Feb 2023 Official leaderboard
15 Curió 1.1B 1.1B 20% Official leaderboard
16 TinyLlama 1T 1.1B 20% Official leaderboard
17 TinyLlama 3T 1.1B 19% Official leaderboard
18 Tucano 1.1B 1.1B 19% Official leaderboard
19 Tucano 2.4B 2.4B 19% Official leaderboard
#ModelScoreSettingDateSource
1 Claude 3.5 Haiku 95% Oct 2024 Official leaderboard
2 Qwen2.5-14B 93% Sep 2024 Official leaderboard
3 Qwen2.5-7B 83% Sep 2024 Official leaderboard
4 Llama 3-8B 61% Apr 2024 Official leaderboard
5 Qwen2.5-3B 59% Sep 2024 Official leaderboard
6 Qwen2.5-1.5B 49% Sep 2024 Official leaderboard
7 Llama 2-7B 23% Jul 2023 Official leaderboard
8 LLaMA-7B 22% Feb 2023 Official leaderboard

Rankings by part

We read 2 parts from the benchmark’s paper: Proverb to Story and Story to Proverb. Each ranks models on one part of the test. Claude 3.5 Haiku leads Proverb to Story and Qwen2.5-14B leads Story to Proverb. Arrows show how far a model moved from its overall rank among the models we track.

Proverb to Story

8 models we track have a score here.

  1. 1 Claude 3.5 Haiku 95%
  2. 2 Qwen2.5-14B 92%
  3. 3 Qwen2.5-7B 83%
  4. 4 Qwen2.5-3B ↑1 51%
  5. 5 Llama 3-8B ↓1 48%
  6. 6 Qwen2.5-1.5B 41%
  7. 7 Llama 2-7B 22%
  8. 8 LLaMA-7B 21%

Number one over time

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

Anthropic
Benchmark paper Passed 85% in Sep 2025, by Claude 3.5 Haiku

Source: The BRoverbs paper

Qwen2.5-14B on the official leaderboard passed 85% in September 2024

Models released before it came out already scored that high. 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% 2023 2024 2025 Near the ceiling (85%) Claude 3.5 Haiku Llama 3-8B Llama 2-7B LLaMA-7B 0% 20% 40% 60% 80% 100% 2023 2024 2025 Near the ceiling (85%) Claude 3.5 Haiku Llama 3-8B

Source: Official BRoverbs 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 Multilingual

All 545 in Multilingual
  • MMMLU In use MMLU professionally translated into 14 languages. Multiple-choice accuracy. Claude Mythos Preview 92.7%
  • WMT24++ In use English-to-55-language translation of WMT24 test data. Scored with automatic metrics such as MetricX or xCOMET. K-EXAONE 90.5%
  • MMLU-ProX In use MMLU-Pro translated into up to 29 languages with expert review. Multiple-choice accuracy per language. Qwen3.7-Max 87%
  • CLIcK In use Models answer 1,995 multiple-choice questions on Korean language and culture. Questions come from official Korean exams and textbooks in eleven categories. Solar Open 2 90.7%
  • Global MMLU Lite 3.0 In use Models answer 400 human-translated MMLU questions per language in 23 languages. Version 3.0 adds Czech, Hungarian, Oriya, Slovak and Tajik and updates Italian. Accuracy. Claude Opus 5 95.7%
  • MMMB In use Models answer multiple-choice questions about images in six languages across 15 categories. The set has about 12,000 questions. LFM2.5-VL-3B 83%

Data audit

All 8 checks pass. Every benchmark

Catalog record 2 pass
  • Pass Task access All tasks are public.
  • Pass Confidence Checked against a primary source.
Scores 2 pass 3 notes
  • Pass Independent runs Scores from the benchmark’s paper (19 entries).
  • Pass Models scored 8 models we track, plus 11 other entries on its leaderboard.
  • Note Newest score The newest came 2 years ago, in October 2024, for Claude 3.5 Haiku. That’s expected for a saturated test.
  • Note Frontier models None of the top 10 has a score, which is fine for a retired test. The strongest tested is Claude 3.5 Haiku (#175).
  • Note Headroom Saturated. The best is 95%, past 85% since September 2024.
Leaderboard import 3 pass
  • Pass Importer We read the board on arxiv.org: 8 rows for models we track and 11 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