Benchmark Almanac

BertaQA

Claude 3 Opus 81.5% leads since June 2024
Not in model cards

Models answer questions about Basque culture and general knowledge. Parallel Basque and English questions help distinguish language effects from familiarity with local topics.

Released
June 2024
Kind
Trivia and open questions
Level
Everyday
Scored
Checked automatically
Answers
Multiple choice
Tasks
All public
Measures
Capability
Language
Basque, English
In and out
Text → Text
Interaction
Single turn
Refresh
One fixed set
Run by
Anyone
Strongest model tested
Claude 3 Opus #173 on the capability index
Last model added
June 2024
Scores refreshed
November 20th, 2026
39
Entries in its paper, 26 models we track

Claude 3 Opus leads with 81.5% among models released by June 2024 Claude 3 Opus leads with 81.5% among models released by June 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. This view shows only models we track, each at its best.

The newest model with a score here came out in June 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 Opus 5-shot 81.5%
  2. 2 GPT-4 Turbo (Jan 2024) 5-shot 80.7%
  3. 3 GPT-4 Jun 2023, 5-shot 77.6%
  4. 4 Meta-Llama-3-70B-Instruct 5-shot 72.4%
  5. 5 Claude 3 Sonnet 5-shot 71.1%
  6. 6 Llama 3-70B 5-shot 71%
  7. 7 Claude 3 Haiku 5-shot 70.3%
  8. 8 latxa-70b-v1.1 5-shot 68.4%
  9. 9 latxa-70b-v1 5-shot 65.9%
  10. 10 Qwen2-72B 5-shot 65.9%
  11. 11 Gemma 2 27B 5-shot 64.3%
  12. 12 gemma-2-27b-it 5-shot 63.5%
  13. 13 GPT-3.5 Turbo Jan 2024, 5-shot 62.8%
  14. 14 latxa-13b-v1.1 5-shot 62.5%
  15. 15 Gemma 2 9B 5-shot 62.4%
  16. 16 gemma-2-9b-it 5-shot 62.1%
  17. 17 Mixtral 8x7B 5-shot 61.3%
  18. 18 Qwen1.5-72B 5-shot 61.3%
  19. 19 latxa-13b-v1 5-shot 60.6%
  20. 20 Yi-34B 5-shot 59.9%
  21. 21 Meta-Llama-3-8B-Instruct 5-shot 59.3%
  22. 22 Llama 3-8B 5-shot 58.2%
  23. 23 Gemma 7B 5-shot 57.5%
  24. 24 Qwen2-7B 5-shot 55%
  25. 25 Llama 2-70B 5-shot 54.7%
  26. 26 Qwen2-7B-Instruct 5-shot 54.3%
  27. 27 Qwen1.5-14B 5-shot 52.9%
  28. 28 Mistral 7B v0.1 5-shot 52.6%
  29. 29 latxa-7b-v1 5-shot 52.2%
  30. 30 Yi-9B 5-shot 51.6%
  31. 31 Yi 6B 5-shot 50.5%
  32. 32 latxa-7b-v1.1 5-shot 50.3%
  33. 33 Qwen1.5-7B 5-shot 49.1%
  34. 34 Llama 2-13B 5-shot 48%
  35. 35 Llama 2-7B 5-shot 44.5%
  36. 36 Gemma 2B 5-shot 41.8%
  37. 37 mGPT-13B 5-shot 34.6%
  38. 38 bloom-7b1 5-shot 34%
  39. 39 xglm-7.5B 5-shot 33.8%

Source: The BertaQA paper

#EntryScoreSettingDateSource
1 Claude 3 Opus 81.5% 5-shot Feb 2024 Official leaderboard
2 GPT-4 Turbo (Jan 2024) 80.7% 5-shot Jan 2024 Official leaderboard
3 GPT-4 Jun 2023 77.6% 5-shot Jun 2023 Official leaderboard
4 Meta-Llama-3-70B-Instruct 72.4% 5-shot Official leaderboard
5 Claude 3 Sonnet 71.1% 5-shot Feb 2024 Official leaderboard
6 Llama 3-70B 71% 5-shot Apr 2024 Official leaderboard
7 Claude 3 Haiku 70.3% 5-shot Mar 2024 Official leaderboard
8 latxa-70b-v1.1 68.4% 5-shot Official leaderboard
9 latxa-70b-v1 65.9% 5-shot Official leaderboard
10 Qwen2-72B 65.9% 5-shot Jun 2024 Official leaderboard
11 Gemma 2 27B 64.3% 5-shot Jun 2024 Official leaderboard
12 gemma-2-27b-it 63.5% 5-shot Official leaderboard
13 GPT-3.5 Turbo Jan 2024 62.8% 5-shot Jan 2024 Official leaderboard
14 latxa-13b-v1.1 62.5% 5-shot Official leaderboard
15 Gemma 2 9B 62.4% 5-shot Jun 2024 Official leaderboard
16 gemma-2-9b-it 62.1% 5-shot Official leaderboard
17 Mixtral 8x7B 61.3% 5-shot Dec 2023 Official leaderboard
18 Qwen1.5-72B 61.3% 5-shot Feb 2024 Official leaderboard
19 latxa-13b-v1 60.6% 5-shot Official leaderboard
20 Yi-34B 59.9% 5-shot Nov 2023 Official leaderboard
21 Meta-Llama-3-8B-Instruct 59.3% 5-shot Official leaderboard
22 Llama 3-8B 58.2% 5-shot Apr 2024 Official leaderboard
23 Gemma 7B 57.5% 5-shot Feb 2024 Official leaderboard
24 Qwen2-7B 55% 5-shot Jun 2024 Official leaderboard
25 Llama 2-70B 54.7% 5-shot Jul 2023 Official leaderboard
26 Qwen2-7B-Instruct 54.3% 5-shot Jun 2024 Official leaderboard
27 Qwen1.5-14B 52.9% 5-shot Feb 2024 Official leaderboard
28 Mistral 7B v0.1 52.6% 5-shot Sep 2023 Official leaderboard
29 latxa-7b-v1 52.2% 5-shot Official leaderboard
30 Yi-9B 51.6% 5-shot Mar 2024 Official leaderboard
31 Yi 6B 50.5% 5-shot Nov 2023 Official leaderboard
32 latxa-7b-v1.1 50.3% 5-shot Official leaderboard
33 Qwen1.5-7B 49.1% 5-shot Feb 2024 Official leaderboard
34 Llama 2-13B 48% 5-shot Jul 2023 Official leaderboard
35 Llama 2-7B 44.5% 5-shot Jul 2023 Official leaderboard
36 Gemma 2B 41.8% 5-shot Feb 2024 Official leaderboard
37 mGPT-13B 34.6% 5-shot Official leaderboard
38 bloom-7b1 34% 5-shot Official leaderboard
39 xglm-7.5B 33.8% 5-shot Official leaderboard
#ModelScoreSettingDateSource
1 Claude 3 Opus 81.5% 5-shot Feb 2024 Official leaderboard
2 GPT-4 Turbo (Jan 2024) 80.7% 5-shot Jan 2024 Official leaderboard
3 GPT-4 Jun 2023 77.6% 5-shot Jun 2023 Official leaderboard
4 Claude 3 Sonnet 71.1% 5-shot Feb 2024 Official leaderboard
5 Llama 3-70B 71% 5-shot Apr 2024 Official leaderboard
6 Claude 3 Haiku 70.3% 5-shot Mar 2024 Official leaderboard
7 Qwen2-72B 65.9% 5-shot Jun 2024 Official leaderboard
8 Gemma 2 27B 64.3% 5-shot Jun 2024 Official leaderboard
9 GPT-3.5 Turbo Jan 2024 62.8% 5-shot Jan 2024 Official leaderboard
10 Gemma 2 9B 62.4% 5-shot Jun 2024 Official leaderboard
11 Mixtral 8x7B 61.3% 5-shot Dec 2023 Official leaderboard
12 Qwen1.5-72B 61.3% 5-shot Feb 2024 Official leaderboard
13 Yi-34B 59.9% 5-shot Nov 2023 Official leaderboard
14 Llama 3-8B 58.2% 5-shot Apr 2024 Official leaderboard
15 Gemma 7B 57.5% 5-shot Feb 2024 Official leaderboard
16 Qwen2-7B 55% 5-shot Jun 2024 Official leaderboard
17 Llama 2-70B 54.7% 5-shot Jul 2023 Official leaderboard
18 Qwen2-7B-Instruct 54.3% 5-shot Jun 2024 Official leaderboard
19 Qwen1.5-14B 52.9% 5-shot Feb 2024 Official leaderboard
20 Mistral 7B v0.1 52.6% 5-shot Sep 2023 Official leaderboard
21 Yi-9B 51.6% 5-shot Mar 2024 Official leaderboard
22 Yi 6B 50.5% 5-shot Nov 2023 Official leaderboard
23 Qwen1.5-7B 49.1% 5-shot Feb 2024 Official leaderboard
24 Llama 2-13B 48% 5-shot Jul 2023 Official leaderboard
25 Llama 2-7B 44.5% 5-shot Jul 2023 Official leaderboard
26 Gemma 2B 41.8% 5-shot Feb 2024 Official leaderboard

Rankings by part

We read 4 parts from the benchmark’s paper, each ranking models on one part of the test, such as English, Basque and English standard error. GPT-4 Turbo (Jan 2024) leads 1 of them, and 3 other models lead the rest. Arrows show how far a model moved from its overall rank among the models we track.

English

26 models we track have a score here.

  1. 1 GPT-4 Turbo (Jan 2024) ↑1 82%
  2. 2 Claude 3 Opus ↓1 81.9%
  3. 3 GPT-4 Jun 2023 80.7%
  4. 4 Qwen2-72B ↑3 73.2%
  5. 5 Claude 3 Sonnet ↓1 72.5%
  6. 6 Llama 3-70B ↓1 72.2%
  7. 7 Claude 3 Haiku ↓1 71.5%
  8. 8 Mixtral 8x7B ↑3 70.2%
  9. 9 Qwen1.5-72B ↑3 69.4%
  10. 10 Yi-34B ↑3 68.9%

Number one over time

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

Anthropic
Benchmark paper Not tested since Jun 2024

Source: The BertaQA 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% 2023 2024 Near the ceiling (85%) Claude 3 Opus GPT-4 0% 20% 40% 60% 80% 100% 2023 2024 Near the ceiling (85%) Claude 3 Opus

Source: Official BertaQA 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 June 2024, for Gemma 2 27B. 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 (39 entries).
  • Pass Models scored 26 models we track, plus 13 other entries on its leaderboard.
  • Warning Newest score The newest came 2 years ago, in June 2024, for Gemma 2 27B.
  • Note Frontier models None of the top 10 has a score. The strongest tested is Claude 3 Opus (#173).
  • Pass Headroom The best is 81.5%, under 85%, so it still separates models.
Leaderboard import 3 pass
  • Pass Importer We read the board on github.com: 27 rows for models we track and 13 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