Benchmark Almanac

EusExams

BSC-LT/ALIA-40b-instruct-2601 62.8% tops its leaderboard Qwen3-8B-Base 51.3% leads the models we track
Not in model cards

Models answer past public service examination questions in Basque. The questions come from several Basque institutions.

Released
March 2024
Kind
Knowledge in many languages
Level
University
Scored
Checked automatically
Answers
Multiple choice
Tasks
All public
Measures
Capability
Language
Basque
In and out
Text → Text
Interaction
Single turn
Items from
Exams and contests
Refresh
One fixed set
Run by
Anyone
Strongest model tested
Qwen3-8B #154 on the capability index
Last model added
April 2025
Scores refreshed
November 20th, 2026
69
Entries on La Leaderboard’s leaderboard, 24 models we track

BSC-LT/ALIA-40b-instruct-2601 tops the leaderboard with 62.8% Qwen3-8B-Base leads with 51.3% among models released by April 2025

Best score per entry from La Leaderboard’s leaderboard, read on November 20th, 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. The best model we track is Qwen3-8B-Base, at 51.3%. This view shows only models we track, each at its best.

The newest model with a score here came out in April 2025, and nobody has tested a newer one. Among the 20 highest-ranked models on the capability index, these have no score yet:

69 entries. Scroll sideways for the rest.

  1. 62.8 BSC-LT/ALIA-40b-instruct-2601
  2. 51.3 Qwen3-8B-Base
  3. 50.8 Gemma 2 9B
  4. 49.4 google/gemma-2-9b-it
  5. 46.6 Gemma 3 4B PT
  6. 46.0 Gemma 3 4B
  7. 45.8 Qwen/Qwen2.5-14B-Instruct-GPTQ-Int8
  8. 44.9 Llama 3.1-8B
  9. 44.8 meta-llama/Meta-Llama-3.1-8B-Instruct
  10. 40.7 BSC-LT/salamandra-7b-instruct
  11. 39.8 Qwen2.5-7B
  12. 39.4 IIC/RigoChat-7b-v2
  13. 39.1 Qwen/Qwen2.5-7B-Instruct
  14. 38.3 BSC-LT/salamandra-7b
  15. 38.2 meta-llama/Llama-3.2-3B-Instruct
  16. 37.6 Llama 3.2 3B
  17. 35.8 Gemma 2 2B IT
  18. 34.2 Aya Expanse 8B
  19. 34.1 Qwen2.5-3B
  20. 34.0 google/gemma-2-2b
  21. 34.0 01-ai/Yi-1.5-9B
  22. 33.5 Qwen/Qwen2.5-3B-Instruct
  23. 33.0 HiTZ/latxa-7b-v1.2
  24. 32.8 Yi-1.5-9B-Chat
  25. 32.8 HuggingFaceTB/SmolLM3-3B-Base
  26. 32.0 occiglot/occiglot-7b-eu5
  27. 31.9 occiglot/occiglot-7b-eu5-instruct
  28. 31.9 occiglot/occiglot-7b-es-en
  29. 31.7 Mistral 7B v0.3
  30. 31.5 mistralai/Mistral-7B-Instruct-v0.3
  31. 31.4 occiglot/occiglot-7b-es-en-instruct
  32. 31.2 utter-project/EuroLLM-9B
  33. 31.1 openGPT-X/Teuken-7B-instruct-commercial-v0.4
  34. 31.0 Qwen2.5-1.5B
  35. 30.9 Qwen3-4B
  36. 30.8 HuggingFaceTB/SmolLM3-3B
  37. 30.5 utter-project/EuroLLM-9B-Instruct
  38. 30.4 Qwen/Qwen2.5-1.5B-Instruct
  39. 29.8 allenai/OLMo-2-1124-7B
  40. 29.8 meta-llama/Llama-3.2-1B-Instruct
  41. 29.2 Gemma 3 1B
  42. 29.1 SmolLM2-1.7B
  43. 29.1 bertin-project/Gromenauer-7B-Instruct
  44. 29.1 allenai/OLMo-2-1124-7B-Instruct
  45. 29.1 openGPT-X/Teuken-7B-instruct-research-v0.4
  46. 28.1 SmolLM2 1.7B Instruct
  47. 28.0 bertin-project/Gromenauer-7B
  48. 27.8 LenguajeNaturalAI/leniachat-qwen2-1.5B-v0
  49. 27.5 Llama 3.2 1B
  50. 26.2 projecte-aina/FLOR-1.3B
  51. 26.0 proxectonos/Carballo-bloom-1.3B
  52. 26.0 utter-project/EuroLLM-1.7B
  53. 26.0 Gemma 3 1B PT
  54. 25.8 utter-project/EuroLLM-1.7B-Instruct
  55. 25.8 unsloth/DeepSeek-R1-Distill-Qwen-14B-bnb-4bit
  56. 25.6 LenguajeNaturalAI/leniachat-gemma-2b-v0
  57. 25.6 Phi-1.5
  58. 25.6 projecte-aina/aguila-7b
  59. 25.6 Qwen3-8B
  60. 25.5 projecte-aina/FLOR-6.3B-Instructed
  61. 25.5 gplsi/Aitana-6.3B
  62. 25.3 bertin-project/bertin-gpt-j-6B
  63. 25.3 projecte-aina/FLOR-1.3B-Instructed
  64. 25.1 projecte-aina/FLOR-6.3B
  65. 25.0 DeepSeek-R1-Distill-Qwen-1.5B
  66. 24.9 BSC-LT/salamandra-2b-instruct
  67. 24.8 DeepSeek-R1-Distill-Qwen-7B
  68. 24.6 Qwen3-1.7B
  69. 23.7 BSC-LT/salamandra-2b
  1. 1 BSC-LT/ALIA-40b-instruct-2601 4-shot 62.8%
  2. 2 Qwen3-8B-Base 51.3%
  3. 3 Gemma 2 9B 4-shot 50.8%
  4. 4 google/gemma-2-9b-it 4-shot 49.4%
  5. 5 Gemma 3 4B PT 46.6%
  6. 6 Gemma 3 4B 46%
  7. 7 Qwen/Qwen2.5-14B-Instruct-GPTQ-Int8 45.8%
  8. 8 Llama 3.1-8B 4-shot 44.9%
  9. 9 meta-llama/Meta-Llama-3.1-8B-Instruct 4-shot 44.8%
  10. 10 BSC-LT/salamandra-7b-instruct 4-shot 40.7%
  11. 11 Qwen2.5-7B 4-shot 39.8%
  12. 12 IIC/RigoChat-7b-v2 4-shot 39.4%
  13. 13 Qwen/Qwen2.5-7B-Instruct 4-shot 39.1%
  14. 14 BSC-LT/salamandra-7b 4-shot 38.3%
  15. 15 meta-llama/Llama-3.2-3B-Instruct 38.2%
  16. 16 Llama 3.2 3B 37.6%
  17. 17 Gemma 2 2B IT 4-shot 35.8%
  18. 18 Aya Expanse 8B 4-shot 34.2%
  19. 19 Qwen2.5-3B 34.1%
  20. 20 google/gemma-2-2b 4-shot 34%
  21. 21 01-ai/Yi-1.5-9B 4-shot 34%
  22. 22 Qwen/Qwen2.5-3B-Instruct 33.5%
  23. 23 HiTZ/latxa-7b-v1.2 4-shot 33%
  24. 24 Yi-1.5-9B-Chat 4-shot 32.8%
  25. 25 HuggingFaceTB/SmolLM3-3B-Base 32.8%
  26. 26 occiglot/occiglot-7b-eu5 4-shot 32%
  27. 27 occiglot/occiglot-7b-eu5-instruct 4-shot 31.9%
  28. 28 occiglot/occiglot-7b-es-en 4-shot 31.9%
  29. 29 Mistral 7B v0.3 4-shot 31.7%
  30. 30 mistralai/Mistral-7B-Instruct-v0.3 4-shot 31.5%
  31. 31 occiglot/occiglot-7b-es-en-instruct 4-shot 31.4%
  32. 32 utter-project/EuroLLM-9B 4-shot 31.2%
  33. 33 openGPT-X/Teuken-7B-instruct-commercial-v0.4 31.1%
  34. 34 Qwen2.5-1.5B 4-shot 31%
  35. 35 Qwen3-4B 30.9%
  36. 36 HuggingFaceTB/SmolLM3-3B 30.8%
  37. 37 utter-project/EuroLLM-9B-Instruct 4-shot 30.5%
  38. 38 Qwen/Qwen2.5-1.5B-Instruct 4-shot 30.4%
  39. 39 allenai/OLMo-2-1124-7B 29.8%
  40. 40 meta-llama/Llama-3.2-1B-Instruct 4-shot 29.8%
  41. 41 Gemma 3 1B 29.2%
  42. 42 SmolLM2-1.7B 29.1%
  43. 43 bertin-project/Gromenauer-7B-Instruct 4-shot 29.1%
  44. 44 allenai/OLMo-2-1124-7B-Instruct 29.1%
  45. 45 openGPT-X/Teuken-7B-instruct-research-v0.4 29.1%
  46. 46 SmolLM2 1.7B Instruct 28.1%
  47. 47 bertin-project/Gromenauer-7B 4-shot 28%
  48. 48 LenguajeNaturalAI/leniachat-qwen2-1.5B-v0 4-shot 27.8%
  49. 49 Llama 3.2 1B 4-shot 27.5%
  50. 50 projecte-aina/FLOR-1.3B 4-shot 26.2%
  51. 51 proxectonos/Carballo-bloom-1.3B 4-shot 26%
  52. 52 utter-project/EuroLLM-1.7B 4-shot 26%
  53. 53 Gemma 3 1B PT 26%
  54. 54 utter-project/EuroLLM-1.7B-Instruct 4-shot 25.8%
  55. 55 unsloth/DeepSeek-R1-Distill-Qwen-14B-bnb-4bit 4-shot 25.8%
  56. 56 LenguajeNaturalAI/leniachat-gemma-2b-v0 4-shot 25.6%
  57. 57 Phi-1.5 4-shot 25.6%
  58. 58 projecte-aina/aguila-7b 4-shot 25.6%
  59. 59 Qwen3-8B 25.6%
  60. 60 projecte-aina/FLOR-6.3B-Instructed 4-shot 25.5%
  61. 61 gplsi/Aitana-6.3B 4-shot 25.5%
  62. 62 bertin-project/bertin-gpt-j-6B 4-shot 25.3%
  63. 63 projecte-aina/FLOR-1.3B-Instructed 4-shot 25.3%
  64. 64 projecte-aina/FLOR-6.3B 4-shot 25.1%
  65. 65 DeepSeek-R1-Distill-Qwen-1.5B 4-shot 25%
  66. 66 BSC-LT/salamandra-2b-instruct 4-shot 24.9%
  67. 67 DeepSeek-R1-Distill-Qwen-7B 4-shot 24.8%
  68. 68 Qwen3-1.7B 24.6%
  69. 69 BSC-LT/salamandra-2b 4-shot 23.7%

Source: La Leaderboard’s leaderboard

#EntryScoreSettingDateSource
1 BSC-LT/ALIA-40b-instruct-2601 62.8% 4-shot Official leaderboard
2 Qwen3-8B-Base 51.3% Apr 2025 Official leaderboard
3 Gemma 2 9B 50.8% 4-shot Jun 2024 Official leaderboard
4 google/gemma-2-9b-it 49.4% 4-shot Official leaderboard
5 Gemma 3 4B PT 46.6% Mar 2025 Official leaderboard
6 Gemma 3 4B 46% Mar 2025 Official leaderboard
7 Qwen/Qwen2.5-14B-Instruct-GPTQ-Int8 45.8% Official leaderboard
8 Llama 3.1-8B 44.9% 4-shot Jul 2024 Official leaderboard
9 meta-llama/Meta-Llama-3.1-8B-Instruct 44.8% 4-shot Official leaderboard
10 BSC-LT/salamandra-7b-instruct 40.7% 4-shot Official leaderboard
11 Qwen2.5-7B 39.8% 4-shot Sep 2024 Official leaderboard
12 IIC/RigoChat-7b-v2 39.4% 4-shot Official leaderboard
13 Qwen/Qwen2.5-7B-Instruct 39.1% 4-shot Official leaderboard
14 BSC-LT/salamandra-7b 38.3% 4-shot Official leaderboard
15 meta-llama/Llama-3.2-3B-Instruct 38.2% Official leaderboard
16 Llama 3.2 3B 37.6% Sep 2024 Official leaderboard
17 Gemma 2 2B IT 35.8% 4-shot Jul 2024 Official leaderboard
18 Aya Expanse 8B 34.2% 4-shot Oct 2024 Official leaderboard
19 Qwen2.5-3B 34.1% Sep 2024 Official leaderboard
20 google/gemma-2-2b 34% 4-shot Official leaderboard
21 01-ai/Yi-1.5-9B 34% 4-shot Official leaderboard
22 Qwen/Qwen2.5-3B-Instruct 33.5% Official leaderboard
23 HiTZ/latxa-7b-v1.2 33% 4-shot Official leaderboard
24 Yi-1.5-9B-Chat 32.8% 4-shot May 2024 Official leaderboard
25 HuggingFaceTB/SmolLM3-3B-Base 32.8% Official leaderboard
26 occiglot/occiglot-7b-eu5 32% 4-shot Official leaderboard
27 occiglot/occiglot-7b-eu5-instruct 31.9% 4-shot Official leaderboard
28 occiglot/occiglot-7b-es-en 31.9% 4-shot Official leaderboard
29 Mistral 7B v0.3 31.7% 4-shot Sep 2023 Official leaderboard
30 mistralai/Mistral-7B-Instruct-v0.3 31.5% 4-shot Official leaderboard
31 occiglot/occiglot-7b-es-en-instruct 31.4% 4-shot Official leaderboard
32 utter-project/EuroLLM-9B 31.2% 4-shot Official leaderboard
33 openGPT-X/Teuken-7B-instruct-commercial-v0.4 31.1% Official leaderboard
34 Qwen2.5-1.5B 31% 4-shot Sep 2024 Official leaderboard
35 Qwen3-4B 30.9% Apr 2025 Official leaderboard
36 HuggingFaceTB/SmolLM3-3B 30.8% Official leaderboard
37 utter-project/EuroLLM-9B-Instruct 30.5% 4-shot Official leaderboard
38 Qwen/Qwen2.5-1.5B-Instruct 30.4% 4-shot Official leaderboard
39 allenai/OLMo-2-1124-7B 29.8% Official leaderboard
40 meta-llama/Llama-3.2-1B-Instruct 29.8% 4-shot Official leaderboard
41 Gemma 3 1B 29.2% Mar 2025 Official leaderboard
42 SmolLM2-1.7B 29.1% Oct 2024 Official leaderboard
43 bertin-project/Gromenauer-7B-Instruct 29.1% 4-shot Official leaderboard
44 allenai/OLMo-2-1124-7B-Instruct 29.1% Official leaderboard
45 openGPT-X/Teuken-7B-instruct-research-v0.4 29.1% Official leaderboard
46 SmolLM2 1.7B Instruct 28.1% Oct 2024 Official leaderboard
47 bertin-project/Gromenauer-7B 28% 4-shot Official leaderboard
48 LenguajeNaturalAI/leniachat-qwen2-1.5B-v0 27.8% 4-shot Official leaderboard
49 Llama 3.2 1B 27.5% 4-shot Sep 2024 Official leaderboard
50 projecte-aina/FLOR-1.3B 26.2% 4-shot Official leaderboard
51 proxectonos/Carballo-bloom-1.3B 26% 4-shot Official leaderboard
52 utter-project/EuroLLM-1.7B 26% 4-shot Official leaderboard
53 Gemma 3 1B PT 26% Mar 2025 Official leaderboard
54 utter-project/EuroLLM-1.7B-Instruct 25.8% 4-shot Official leaderboard
55 unsloth/DeepSeek-R1-Distill-Qwen-14B-bnb-4bit 25.8% 4-shot Official leaderboard
56 LenguajeNaturalAI/leniachat-gemma-2b-v0 25.6% 4-shot Official leaderboard
57 Phi-1.5 25.6% 4-shot Sep 2023 Official leaderboard
58 projecte-aina/aguila-7b 25.6% 4-shot Official leaderboard
59 Qwen3-8B 25.6% Apr 2025 Official leaderboard
60 projecte-aina/FLOR-6.3B-Instructed 25.5% 4-shot Official leaderboard
61 gplsi/Aitana-6.3B 25.5% 4-shot Official leaderboard
62 bertin-project/bertin-gpt-j-6B 25.3% 4-shot Official leaderboard
63 projecte-aina/FLOR-1.3B-Instructed 25.3% 4-shot Official leaderboard
64 projecte-aina/FLOR-6.3B 25.1% 4-shot Official leaderboard
65 DeepSeek-R1-Distill-Qwen-1.5B 25% 4-shot Jan 2025 Official leaderboard
66 BSC-LT/salamandra-2b-instruct 24.9% 4-shot Official leaderboard
67 DeepSeek-R1-Distill-Qwen-7B 24.8% 4-shot Jan 2025 Official leaderboard
68 Qwen3-1.7B 24.6% Apr 2025 Official leaderboard
69 BSC-LT/salamandra-2b 23.7% 4-shot Official leaderboard
#ModelScoreSettingDateSource
1 Qwen3-8B-Base 51.3% Apr 2025 Official leaderboard
2 Gemma 2 9B 50.8% 4-shot Jun 2024 Official leaderboard
3 Gemma 3 4B PT 46.6% Mar 2025 Official leaderboard
4 Gemma 3 4B 46% Mar 2025 Official leaderboard
5 Llama 3.1-8B 44.9% 4-shot Jul 2024 Official leaderboard
6 Qwen2.5-7B 39.8% 4-shot Sep 2024 Official leaderboard
7 Llama 3.2 3B 37.6% Sep 2024 Official leaderboard
8 Gemma 2 2B IT 35.8% 4-shot Jul 2024 Official leaderboard
9 Aya Expanse 8B 34.2% 4-shot Oct 2024 Official leaderboard
10 Qwen2.5-3B 34.1% Sep 2024 Official leaderboard
11 Yi-1.5-9B-Chat 32.8% 4-shot May 2024 Official leaderboard
12 Mistral 7B v0.3 31.7% 4-shot Sep 2023 Official leaderboard
13 Qwen2.5-1.5B 31% 4-shot Sep 2024 Official leaderboard
14 Qwen3-4B 30.9% Apr 2025 Official leaderboard
15 Gemma 3 1B 29.2% Mar 2025 Official leaderboard
16 SmolLM2-1.7B 29.1% Oct 2024 Official leaderboard
17 SmolLM2 1.7B Instruct 28.1% Oct 2024 Official leaderboard
18 Llama 3.2 1B 27.5% 4-shot Sep 2024 Official leaderboard
19 Gemma 3 1B PT 26% Mar 2025 Official leaderboard
20 Phi-1.5 25.6% 4-shot Sep 2023 Official leaderboard
21 Qwen3-8B 25.6% Apr 2025 Official leaderboard
22 DeepSeek-R1-Distill-Qwen-1.5B 25% 4-shot Jan 2025 Official leaderboard
23 DeepSeek-R1-Distill-Qwen-7B 24.8% 4-shot Jan 2025 Official leaderboard
24 Qwen3-1.7B 24.6% Apr 2025 Official leaderboard

Rankings by part

We read 29 parts from La Leaderboard’s leaderboard, each ranking models on one part of the test, such as ejadministrari, ejlaguntza and ejlaguntzaile. Qwen3-8B-Base leads 25 of them, and Gemma 3 4B PT leads the other 4. Arrows show how far a model moved from its overall rank among the models we track.

ejadministrari

12 models we track have a score here.

  1. 1 Qwen3-8B-Base 53.4%
  2. 2 Gemma 3 4B PT ↑1 48.7%
  3. 3 Gemma 3 4B ↑1 48.4%
  4. 4 Llama 3.2 3B ↑3 41.1%
  5. 5 Qwen2.5-3B ↑5 36.2%
  6. 6 Qwen3-4B ↑8 32.2%
  7. 7 Gemma 3 1B ↑8 31.3%
  8. 8 SmolLM2 1.7B Instruct ↑9 30.6%
  9. 9 SmolLM2-1.7B ↑7 29.9%
  10. 10 Gemma 3 1B PT ↑9 26%

4 models have held number one on La Leaderboard’s leaderboard since its release in March 2024

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

Mistral AI 01.AI Google DeepMind Qwen
La Leaderboard leaderboard Not tested since Apr 2025

Source: La Leaderboard’s leaderboard

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
On La Leaderboard’s leaderboard Best so far
0% 20% 40% 60% 80% 100% 2024 2025 Near the ceiling (85%) Qwen3-8B-Base Gemma 2 9B Yi-1.5-9B-Chat Mistral 7B v0.3 Phi-1.5 0% 20% 40% 60% 80% 100% 2024 2025 Near the ceiling (85%) Qwen3-8B-Base Yi-1.5-9B-Chat

Source: La Leaderboard’s 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

9 of 10 checks pass, 1 warning. Look first at the Newest score check. The newest came 19 months ago, in April 2025, for Qwen3-1.7B. 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 official leaderboard (69 entries).
  • Pass Models scored 24 models we track, plus 45 other entries on its leaderboard.
  • Warning Newest score The newest came 19 months ago, in April 2025, for Qwen3-1.7B.
  • Note Frontier models None of the top 10 has a score. The strongest tested is Qwen3-8B (#154).
  • Pass Headroom The best is 51.3%, under 85%, so it still separates models.
Leaderboard import 3 pass
  • Pass Importer We read the board on La Leaderboard: 24 rows for models we track and 45 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