Benchmark Almanac

Miðeind Icelandic Case Government

Claude Fable 5.2 100% leads since November 2026
Not in model cards

Models inflect a subject or construct a verb and object phrase in Icelandic. The verb determines the grammatical case. The evaluator checks exact agreement.

Passes test 2. Miðeind reports scores for 60 named models on its current leaderboard.

Made by
Miðeind
Released
September 2026
Kind
Reasoning in many languages
Level
School
Scored
Checked automatically
Answers
Short answer
Tasks
Kept private
Measures
Capability
Language
Icelandic
In and out
Text → Text
Interaction
Single turn
Items from
Written by experts
Refresh
One fixed set
Run by
A host
Strongest model tested
Claude Fable 5.2 #2 on the capability index
Last model added
November 2026
Scores refreshed
November 20th, 2026
62
Entries on its own leaderboard, 61 models we track

2 models share the top score of 100% 2 models share the top score of 100%

Best score per entry from the benchmark's own leaderboard, read on November 20th, 2026. Where a model ran with several settings or harnesses, we keep its best. Every run is in the next chart. 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.

  1. 1 Claude Fable 5.2 high 100%
  2. 2 Gemini 4 Hydrogen high 100%
  3. 3 Claude Opus 5 high 99.1%
  4. 4 Gemini 3.1 Pro low 98.8%
  5. 5 Gemini 3.8 Flash high 98.8%
  6. 6 Gemini 4 Argon high 98.8%
  7. 7 Gemini 3.7 Flash high 98.4%
  8. 8 Gemini 4 Carbon low 98.4%
  9. 9 Gemini 4 Neon low 98.4%
  10. 10 Gemini 3.5 Flash high 97.5%
  11. 11 Claude Opus 5.5 high 97.5%
  12. 12 GPT-6.1 Terra high 97.5%
  13. 13 GPT-6.2 Sol low 97.5%
  14. 14 GPT-6.1 Sol high 97.2%
  15. 15 GPT-6 Astra high 96.6%
  16. 16 Claude Fable 5.1 high 96.3%
  17. 17 Claude Opus 4.8 high 95.9%
  18. 18 GPT-5.6 Sol high 95.3%
  19. 19 Claude Sonnet 5.5 high 95.3%
  20. 20 GPT-6 Sol high 94.4%
  21. 21 Claude Sonnet 5 high 89.1%
  22. 22 Gemini 3.5 Flash-Lite high 88.1%
  23. 23 Muse Spark Open high 87.8%
  24. 24 Inkling 2 Small high 85.6%
  25. 25 Gemini 2.5 Pro Jun 2025, high 84.1%
  26. 26 Qwen3.8-2.4T-A95B on 80.3%
  27. 27 GPT-6 Luna high 79.4%
  28. 28 GPT-5.2 high 78.4%
  29. 29 DeepSeek-V4-Pro on 78.4%
  30. 30 DeepSeek V4.1 Flash on 76.9%
  31. 31 Grok 4.7 high 76.9%
  32. 32 GPT-5.6 Terra high 75%
  33. 33 Inkling high 71.3%
  34. 34 GPT-5.6 Luna high 70%
  35. 35 Claude Haiku 5.5 high 64.1%
  36. 36 GPT-4o Mar 2025, none 61.6%
  37. 37 Kimi K3 on 55.6%
  38. 38 Mistral Large 4 high 45%
  39. 39 Inkling-Small high 43.1%
  40. 40 GLM-5.3-Flash on 41.6%
  41. 41 MiniMax-M3 on 33.8%
  42. 42 Mistral Large 3 none 26.6%
  43. 43 Mistral Small 4 high 25.3%
  44. 44 Gemini 2.5 Flash-Lite Sep 2025, off 24.4%
  45. 45 Qwen 3.8 27B xhigh 21.3%
  46. 46 Gemma 4 31B IT on 16.6%
  47. 47 Claude Haiku 4.5 off 14.7%
  48. 48 Gemma 4 26B A4B on 14.7%
  49. 49 MiniMax-M3.1-Flash high 10.6%
  50. 50 Llama 3.1-70B none 7.50%
  51. 51 Mistral Medium 3.5 high 6.25%
  52. 52 Bielik 11B v3.1 Instruct high 5.94%
  53. 53 Gemma 4 12B off 5.63%
  54. 54 Ripple-65B Preview on 5.31%
  55. 55 Apertus 1.5 8B on 4.69%
  56. 56 Apertus 1.5 70B on 4.06%
  57. 57 GPT-4.1 nano none 3.75%
  58. 58 Gemma 4 E4B on 3.75%
  59. 59 GPT-3.5 Turbo Jan 2024, none 2.81%
  60. 60 Llama 3.1-8B none 2.19%
  61. 61 Apertus 8B Instruct none 2.19%
  62. 62 Gemma 4 E2B off 2.19%

Source: Official Miðeind Icelandic Case Government leaderboard

  1. 1 Claude Fable 5.2 high 100%
  2. 2 Gemini 4 Hydrogen high 100%
  3. 3 Claude Opus 5 high 99.1%
  4. 4 Gemini 3.1 Pro low 98.8%
  5. 5 Gemini 3.8 Flash high 98.8%
  6. 6 Gemini 4 Argon high 98.8%
  7. 7 Gemini 3.7 Flash high 98.4%
  8. 8 Gemini 4 Carbon low 98.4%
  9. 9 Gemini 4 Neon low 98.4%
  10. 10 Gemini 3.5 Flash high 97.5%
  11. 11 Claude Opus 5.5 high 97.5%
  12. 12 GPT-6.1 Terra high 97.5%
  13. 13 GPT-6.2 Sol low 97.5%
  14. 14 GPT-6.1 Sol high 97.2%
  15. 15 GPT-6 Astra high 96.6%
  16. 16 Claude Fable 5.1 high 96.3%
  17. 17 Claude Opus 4.8 high 95.9%
  18. 18 GPT-5.6 Sol high 95.3%
  19. 19 Claude Sonnet 5.5 high 95.3%
  20. 20 GPT-6 Sol high 94.4%
  21. 21 Claude Sonnet 5 high 89.1%
  22. 22 Gemini 3.5 Flash-Lite high 88.1%
  23. 23 Muse Spark Open high 87.8%
  24. 24 Inkling 2 Small high 85.6%
  25. 25 Gemini 2.5 Pro Jun 2025, high 84.1%
  26. 26 Qwen3.8-2.4T-A95B on 80.3%
  27. 27 GPT-6 Luna high 79.4%
  28. 28 GPT-5.2 high 78.4%
  29. 29 DeepSeek-V4-Pro on 78.4%
  30. 30 DeepSeek V4.1 Flash on 76.9%
  31. 31 Grok 4.7 high 76.9%
  32. 32 GPT-5.6 Terra high 75%
  33. 33 Inkling high 71.3%
  34. 34 GPT-5.6 Luna high 70%
  35. 35 Claude Haiku 5.5 high 64.1%
  36. 36 GPT-4o Mar 2025, none 61.6%
  37. 37 Kimi K3 on 55.6%
  38. 38 Mistral Large 4 high 45%
  39. 39 Inkling-Small high 43.1%
  40. 40 GLM-5.3-Flash on 41.6%
  41. 41 MiniMax-M3 on 33.8%
  42. 42 Mistral Large 3 none 26.6%
  43. 43 Mistral Small 4 high 25.3%
  44. 44 Gemini 2.5 Flash-Lite Sep 2025, off 24.4%
  45. 45 Qwen 3.8 27B xhigh 21.3%
  46. 46 Gemma 4 31B IT on 16.6%
  47. 47 Claude Haiku 4.5 off 14.7%
  48. 48 Gemma 4 26B A4B on 14.7%
  49. 49 MiniMax-M3.1-Flash high 10.6%
  50. 50 Llama 3.1-70B none 7.50%
  51. 51 Mistral Medium 3.5 high 6.25%
  52. 52 Gemma 4 12B off 5.63%
  53. 53 Ripple-65B Preview on 5.31%
  54. 54 Apertus 1.5 8B on 4.69%
  55. 55 Apertus 1.5 70B on 4.06%
  56. 56 GPT-4.1 nano none 3.75%
  57. 57 Gemma 4 E4B on 3.75%
  58. 58 GPT-3.5 Turbo Jan 2024, none 2.81%
  59. 59 Llama 3.1-8B none 2.19%
  60. 60 Apertus 8B Instruct none 2.19%
  61. 61 Gemma 4 E2B off 2.19%

Source: Official Miðeind Icelandic Case Government leaderboard

#EntryScoreSettingReleasedRun onLinkSource
1 Claude Fable 5.2 100% high Nov 2026 Nov 2026 anthropic.com Official leaderboard
2 Gemini 4 Hydrogen 100% high Nov 2026 Nov 2026 ai.google.dev Official leaderboard
3 Claude Opus 5 99.1% high Jul 2026 Sep 2026 anthropic.com Official leaderboard
4 Gemini 3.1 Pro 98.8% low Feb 2026 Sep 2026 ai.google.dev Official leaderboard
5 Gemini 3.8 Flash 98.8% high Sep 2026 Sep 2026 ai.google.dev Official leaderboard
6 Gemini 4 Argon 98.8% high Sep 2026 Oct 2026 ai.google.dev Official leaderboard
7 Gemini 3.7 Flash 98.4% high Aug 2026 Sep 2026 ai.google.dev Official leaderboard
8 Gemini 4 Carbon 98.4% low Oct 2026 Nov 2026 ai.google.dev Official leaderboard
9 Gemini 4 Neon 98.4% low Nov 2026 Nov 2026 ai.google.dev Official leaderboard
10 Gemini 3.5 Flash 97.5% high May 2026 Sep 2026 ai.google.dev Official leaderboard
11 Claude Opus 5.5 97.5% high Sep 2026 Sep 2026 anthropic.com Official leaderboard
12 GPT-6.1 Terra 97.5% high Oct 2026 Oct 2026 developers.openai.com Official leaderboard
13 GPT-6.2 Sol 97.5% low Nov 2026 Nov 2026 developers.openai.com Official leaderboard
14 GPT-6.1 Sol 97.2% high Sep 2026 Sep 2026 developers.openai.com Official leaderboard
15 GPT-6 Astra 96.6% high Sep 2026 Sep 2026 developers.openai.com Official leaderboard
16 Claude Fable 5.1 96.3% high Sep 2026 Sep 2026 anthropic.com Official leaderboard
17 Claude Opus 4.8 95.9% high May 2026 Sep 2026 anthropic.com Official leaderboard
18 GPT-5.6 Sol 95.3% high Jul 2026 Sep 2026 developers.openai.com Official leaderboard
19 Claude Sonnet 5.5 95.3% high Sep 2026 Sep 2026 anthropic.com Official leaderboard
20 GPT-6 Sol 94.4% high Sep 2026 Sep 2026 developers.openai.com Official leaderboard
21 Claude Sonnet 5 89.1% high Jun 2026 Sep 2026 anthropic.com Official leaderboard
22 Gemini 3.5 Flash-Lite 88.1% high Jul 2026 Sep 2026 ai.google.dev Official leaderboard
23 Muse Spark Open 87.8% high Oct 2026 Oct 2026 huggingface.co Official leaderboard
24 Inkling 2 Small 85.6% high Nov 2026 Nov 2026 huggingface.co Official leaderboard
25 Gemini 2.5 Pro Jun 2025 84.1% high Jun 2025 Sep 2026 ai.google.dev Official leaderboard
26 Qwen3.8-2.4T-A95B 80.3% on Aug 2026 Sep 2026 huggingface.co Official leaderboard
27 GPT-6 Luna 79.4% high Sep 2026 Sep 2026 developers.openai.com Official leaderboard
28 GPT-5.2 78.4% high Dec 2025 Sep 2026 platform.openai.com Official leaderboard
29 DeepSeek-V4-Pro 78.4% on Apr 2026 Sep 2026 huggingface.co Official leaderboard
30 DeepSeek V4.1 Flash 76.9% on Sep 2026 Oct 2026 huggingface.co Official leaderboard
31 Grok 4.7 76.9% high Sep 2026 Nov 2026 docs.x.ai Official leaderboard
32 GPT-5.6 Terra 75% high Jul 2026 Sep 2026 developers.openai.com Official leaderboard
33 Inkling 71.3% high Jul 2026 Sep 2026 huggingface.co Official leaderboard
34 GPT-5.6 Luna 70% high Jul 2026 Sep 2026 developers.openai.com Official leaderboard
35 Claude Haiku 5.5 64.1% high Oct 2026 Oct 2026 anthropic.com Official leaderboard
36 GPT-4o Mar 2025 61.6% none Mar 2025 Sep 2026 platform.openai.com Official leaderboard
37 Kimi K3 55.6% on Jul 2026 Sep 2026 huggingface.co Official leaderboard
38 Mistral Large 4 45% high Oct 2026 Oct 2026 mistral.ai Official leaderboard
39 Inkling-Small 43.1% high Jul 2026 Sep 2026 huggingface.co Official leaderboard
40 GLM-5.3-Flash 41.6% on Aug 2026 Sep 2026 huggingface.co Official leaderboard
41 MiniMax-M3 33.8% on Jun 2026 Sep 2026 huggingface.co Official leaderboard
42 Mistral Large 3 26.6% none Dec 2025 Sep 2026 mistral.ai Official leaderboard
43 Mistral Small 4 25.3% high Mar 2026 Sep 2026 mistral.ai Official leaderboard
44 Gemini 2.5 Flash-Lite Sep 2025 24.4% off Sep 2025 Sep 2026 ai.google.dev Official leaderboard
45 Qwen 3.8 27B 21.3% xhigh Aug 2026 Oct 2026 huggingface.co Official leaderboard
46 Gemma 4 31B IT 16.6% on Apr 2026 Sep 2026 deepmind.google Official leaderboard
47 Claude Haiku 4.5 14.7% off Oct 2025 Sep 2026 anthropic.com Official leaderboard
48 Gemma 4 26B A4B 14.7% on Apr 2026 Oct 2026 huggingface.co Official leaderboard
49 MiniMax-M3.1-Flash 10.6% high Oct 2026 Oct 2026 huggingface.co Official leaderboard
50 Llama 3.1-70B 7.50% none Jul 2024 Oct 2026 huggingface.co Official leaderboard
51 Mistral Medium 3.5 6.25% high Apr 2026 Sep 2026 mistral.ai Official leaderboard
52 Bielik 11B v3.1 Instruct 11.2B 5.94% high Nov 2026 huggingface.co Official leaderboard
53 Gemma 4 12B 5.63% off Jun 2026 Nov 2026 huggingface.co Official leaderboard
54 Ripple-65B Preview 5.31% on Nov 2026 Nov 2026 huggingface.co Official leaderboard
55 Apertus 1.5 8B 4.69% on Jul 2026 Sep 2026 huggingface.co Official leaderboard
56 Apertus 1.5 70B 4.06% on Jul 2026 Sep 2026 huggingface.co Official leaderboard
57 GPT-4.1 nano 3.75% none Apr 2025 Sep 2026 platform.openai.com Official leaderboard
58 Gemma 4 E4B 3.75% on Mar 2026 Oct 2026 huggingface.co Official leaderboard
59 GPT-3.5 Turbo Jan 2024 2.81% none Jan 2024 Sep 2026 platform.openai.com Official leaderboard
60 Llama 3.1-8B 2.19% none Jul 2024 Oct 2026 huggingface.co Official leaderboard
61 Apertus 8B Instruct 2.19% none Sep 2025 Sep 2026 huggingface.co Official leaderboard
62 Gemma 4 E2B 2.19% off Mar 2026 Oct 2026 huggingface.co Official leaderboard
#ModelScoreSettingReleasedRun onLinkSource
1 Claude Fable 5.2 100% high Nov 2026 Nov 2026 anthropic.com Official leaderboard
2 Gemini 4 Hydrogen 100% high Nov 2026 Nov 2026 ai.google.dev Official leaderboard
3 Claude Opus 5 99.1% high Jul 2026 Sep 2026 anthropic.com Official leaderboard
4 Gemini 3.1 Pro 98.8% low Feb 2026 Sep 2026 ai.google.dev Official leaderboard
5 Gemini 3.8 Flash 98.8% high Sep 2026 Sep 2026 ai.google.dev Official leaderboard
6 Gemini 4 Argon 98.8% high Sep 2026 Oct 2026 ai.google.dev Official leaderboard
7 Gemini 3.7 Flash 98.4% high Aug 2026 Sep 2026 ai.google.dev Official leaderboard
8 Gemini 4 Carbon 98.4% low Oct 2026 Nov 2026 ai.google.dev Official leaderboard
9 Gemini 4 Neon 98.4% low Nov 2026 Nov 2026 ai.google.dev Official leaderboard
10 Gemini 3.5 Flash 97.5% high May 2026 Sep 2026 ai.google.dev Official leaderboard
11 Claude Opus 5.5 97.5% high Sep 2026 Sep 2026 anthropic.com Official leaderboard
12 GPT-6.1 Terra 97.5% high Oct 2026 Oct 2026 developers.openai.com Official leaderboard
13 GPT-6.2 Sol 97.5% low Nov 2026 Nov 2026 developers.openai.com Official leaderboard
14 GPT-6.1 Sol 97.2% high Sep 2026 Sep 2026 developers.openai.com Official leaderboard
15 GPT-6 Astra 96.6% high Sep 2026 Sep 2026 developers.openai.com Official leaderboard
16 Claude Fable 5.1 96.3% high Sep 2026 Sep 2026 anthropic.com Official leaderboard
17 Claude Opus 4.8 95.9% high May 2026 Sep 2026 anthropic.com Official leaderboard
18 GPT-5.6 Sol 95.3% high Jul 2026 Sep 2026 developers.openai.com Official leaderboard
19 Claude Sonnet 5.5 95.3% high Sep 2026 Sep 2026 anthropic.com Official leaderboard
20 GPT-6 Sol 94.4% high Sep 2026 Sep 2026 developers.openai.com Official leaderboard
21 Claude Sonnet 5 89.1% high Jun 2026 Sep 2026 anthropic.com Official leaderboard
22 Gemini 3.5 Flash-Lite 88.1% high Jul 2026 Sep 2026 ai.google.dev Official leaderboard
23 Muse Spark Open 87.8% high Oct 2026 Oct 2026 huggingface.co Official leaderboard
24 Inkling 2 Small 85.6% high Nov 2026 Nov 2026 huggingface.co Official leaderboard
25 Gemini 2.5 Pro Jun 2025 84.1% high Jun 2025 Sep 2026 ai.google.dev Official leaderboard
26 Qwen3.8-2.4T-A95B 80.3% on Aug 2026 Sep 2026 huggingface.co Official leaderboard
27 GPT-6 Luna 79.4% high Sep 2026 Sep 2026 developers.openai.com Official leaderboard
28 GPT-5.2 78.4% high Dec 2025 Sep 2026 platform.openai.com Official leaderboard
29 DeepSeek-V4-Pro 78.4% on Apr 2026 Sep 2026 huggingface.co Official leaderboard
30 DeepSeek V4.1 Flash 76.9% on Sep 2026 Oct 2026 huggingface.co Official leaderboard
31 Grok 4.7 76.9% high Sep 2026 Nov 2026 docs.x.ai Official leaderboard
32 GPT-5.6 Terra 75% high Jul 2026 Sep 2026 developers.openai.com Official leaderboard
33 Inkling 71.3% high Jul 2026 Sep 2026 huggingface.co Official leaderboard
34 GPT-5.6 Luna 70% high Jul 2026 Sep 2026 developers.openai.com Official leaderboard
35 Claude Haiku 5.5 64.1% high Oct 2026 Oct 2026 anthropic.com Official leaderboard
36 GPT-4o Mar 2025 61.6% none Mar 2025 Sep 2026 platform.openai.com Official leaderboard
37 Kimi K3 55.6% on Jul 2026 Sep 2026 huggingface.co Official leaderboard
38 Mistral Large 4 45% high Oct 2026 Oct 2026 mistral.ai Official leaderboard
39 Inkling-Small 43.1% high Jul 2026 Sep 2026 huggingface.co Official leaderboard
40 GLM-5.3-Flash 41.6% on Aug 2026 Sep 2026 huggingface.co Official leaderboard
41 MiniMax-M3 33.8% on Jun 2026 Sep 2026 huggingface.co Official leaderboard
42 Mistral Large 3 26.6% none Dec 2025 Sep 2026 mistral.ai Official leaderboard
43 Mistral Small 4 25.3% high Mar 2026 Sep 2026 mistral.ai Official leaderboard
44 Gemini 2.5 Flash-Lite Sep 2025 24.4% off Sep 2025 Sep 2026 ai.google.dev Official leaderboard
45 Qwen 3.8 27B 21.3% xhigh Aug 2026 Oct 2026 huggingface.co Official leaderboard
46 Gemma 4 31B IT 16.6% on Apr 2026 Sep 2026 deepmind.google Official leaderboard
47 Claude Haiku 4.5 14.7% off Oct 2025 Sep 2026 anthropic.com Official leaderboard
48 Gemma 4 26B A4B 14.7% on Apr 2026 Oct 2026 huggingface.co Official leaderboard
49 MiniMax-M3.1-Flash 10.6% high Oct 2026 Oct 2026 huggingface.co Official leaderboard
50 Llama 3.1-70B 7.50% none Jul 2024 Oct 2026 huggingface.co Official leaderboard
51 Mistral Medium 3.5 6.25% high Apr 2026 Sep 2026 mistral.ai Official leaderboard
52 Gemma 4 12B 5.63% off Jun 2026 Nov 2026 huggingface.co Official leaderboard
53 Ripple-65B Preview 5.31% on Nov 2026 Nov 2026 huggingface.co Official leaderboard
54 Apertus 1.5 8B 4.69% on Jul 2026 Sep 2026 huggingface.co Official leaderboard
55 Apertus 1.5 70B 4.06% on Jul 2026 Sep 2026 huggingface.co Official leaderboard
56 GPT-4.1 nano 3.75% none Apr 2025 Sep 2026 platform.openai.com Official leaderboard
57 Gemma 4 E4B 3.75% on Mar 2026 Oct 2026 huggingface.co Official leaderboard
58 GPT-3.5 Turbo Jan 2024 2.81% none Jan 2024 Sep 2026 platform.openai.com Official leaderboard
59 Llama 3.1-8B 2.19% none Jul 2024 Oct 2026 huggingface.co Official leaderboard
60 Apertus 8B Instruct 2.19% none Sep 2025 Sep 2026 huggingface.co Official leaderboard
61 Gemma 4 E2B 2.19% off Mar 2026 Oct 2026 huggingface.co Official leaderboard

Rankings by part

We read 2 parts from the benchmark’s own leaderboard: Subject and Construct. Each ranks models on one part of the test. Claude Fable 5.2 leads Subject and Gemini 3.1 Pro leads Construct. Arrows show how far a model moved from its overall rank among the models we track.

Subject

61 models we track have a score here.

  1. 1 Claude Fable 5.2 100%
  2. 2 Gemini 4 Hydrogen 100%
  3. 3 Claude Opus 5 99%
  4. 4 Gemini 3.8 Flash ↑1 99%
  5. 5 Gemini 3.1 Pro ↓1 98.6%
  6. 6 Gemini 4 Argon 98.6%
  7. 7 Gemini 4 Carbon ↑1 98.6%
  8. 8 Gemini 3.7 Flash ↓1 98.3%
  9. 9 Gemini 4 Neon 98.3%
  10. 10 GPT-6.2 Sol ↑3 97.9%

More effort adds a median 2 points

Models the leaderboard ran at two or more effort levels. We don’t have enough cost data here to say what the extra effort costs.

Lowest effort to highest, light to dark
Claude Fable 5.2
low
100%
Gemini 4 Hydrogen
low
100%
Claude Opus 5
off high
99.1%
Gemini 3.1 Pro
low
98.8%
Gemini 3.8 Flash
low high
98.8%
Gemini 4 Argon
off
98.8%
Gemini 3.7 Flash
low
98.4%
Gemini 3.5 Flash
off high
97.5%
GPT-6.1 Terra
off high
97.5%
GPT-6.2 Sol
low
97.5%
GPT-6.1 Sol
low high
97.2%
GPT-6 Astra
low
96.6%

Source: Official Miðeind Icelandic Case Government leaderboard

ModelEffortHarnessScoreCost per taskOutput tokens
Claude Fable 5.2 high 100%
Gemini 4 Hydrogen high 100%
Claude Fable 5.2 low 99.4%
Gemini 4 Hydrogen low 99.4%
Claude Opus 5 high 99.1%
Gemini 3.1 Pro low 98.8%
Gemini 3.8 Flash high 98.8%
Gemini 4 Argon high 98.8%
Gemini 3.7 Flash high 98.4%
Gemini 4 Carbon low 98.4%
Gemini 4 Neon low 98.4%
Gemini 3.1 Pro high 97.8%
Gemini 3.5 Flash high 97.5%
Claude Opus 5.5 high 97.5%
Gemini 4 Argon off 97.5%
GPT-6.1 Terra high 97.5%
GPT-6.2 Sol low 97.5%
GPT-6.1 Sol high 97.2%
GPT-6.2 Sol high 97.2%
Gemini 3.7 Flash low 96.9%
GPT-6 Astra high 96.6%
Claude Fable 5.1 high 96.3%
GPT-6 Astra low 96.3%
Claude Opus 4.8 high 95.9%
Gemini 3.8 Flash low 95.9%
GPT-5.6 Sol high 95.3%
Claude Sonnet 5.5 high 95.3%
GPT-6 Sol high 94.4%
GPT-6.1 Sol low 94.1%
GPT-6.1 Terra off 93.4%
GPT-5.6 Sol off 90.9%
Claude Sonnet 5 high 89.1%
Claude Opus 4.8 off 88.8%
Gemini 3.5 Flash-Lite high 88.1%
Muse Spark Open high 87.8%
GPT-6 Sol off 87.5%
Inkling 2 Small high 85.6%
Gemini 2.5 Pro Jun 2025 high 84.1%
Gemini 3.5 Flash off 83.1%
Claude Sonnet 5 off 82.2%
Inkling 2 Small minimal 82.2%
Gemini 2.5 Pro Jun 2025 minimal 81.3%
Qwen3.8-2.4T-A95B on 80.3%
Claude Sonnet 5.5 off 80%
GPT-6 Luna high 79.4%
GPT-5.2 high 78.4%
DeepSeek-V4-Pro on 78.4%
DeepSeek V4.1 Flash on 76.9%
Grok 4.7 high 76.9%
Grok 4.7 low 75.3%
GPT-5.6 Terra high 75%
GPT-5.6 Terra off 75%
DeepSeek-V4-Pro off 73.8%
Inkling high 71.3%
GPT-5.6 Luna high 70%
Inkling off 69.4%
Claude Opus 5 off 64.4%
Claude Haiku 5.5 high 64.1%
GPT-5.6 Luna off 63.4%
GPT-6 Luna off 62.8%
GPT-4o Mar 2025 none 61.6%
Gemini 3.5 Flash-Lite off 61.6%
Kimi K3 on 55.6%
GPT-5.2 off 53.8%
DeepSeek V4.1 Flash off 47.2%
Mistral Large 4 high 45%
Muse Spark Open off 44.7%
Inkling-Small high 43.1%
GLM-5.3-Flash on 41.6%
MiniMax-M3 on 33.8%
Mistral Large 4 low 33.4%
Kimi K3 off 26.9%
Mistral Large 3 none 26.6%
Claude Haiku 5.5 off 26.6%
Mistral Small 4 high 25.3%
Gemini 2.5 Flash-Lite Sep 2025 off 24.4%
Qwen 3.8 27B xhigh 21.3%
Gemini 2.5 Flash-Lite Sep 2025 high 18.4%
Mistral Small 4 off 17.5%
Gemma 4 31B IT on 16.6%
Claude Haiku 4.5 off 14.7%
Gemma 4 26B A4B on 14.7%
MiniMax-M3.1-Flash high 10.6%
Qwen 3.8 27B off 9.38%
Claude Haiku 4.5 high 7.81%
Llama 3.1-70B none 7.50%
Mistral Medium 3.5 high 6.25%
Mistral Medium 3.5 off 5.63%
Gemma 4 12B off 5.63%
Gemma 4 26B A4B off 5.31%
Gemma 4 12B on 5.31%
Ripple-65B Preview on 5.31%
Apertus 1.5 8B on 4.69%
MiniMax-M3.1-Flash off 4.38%
Apertus 1.5 70B on 4.06%
GPT-4.1 nano none 3.75%
Gemma 4 E4B on 3.75%
Gemma 4 E4B off 3.44%
GPT-3.5 Turbo Jan 2024 none 2.81%
Llama 3.1-8B none 2.19%
Apertus 8B Instruct none 2.19%
Gemma 4 E2B off 2.19%
Gemma 4 E2B on 2.19%

More effort adds a median 2 points

Models the leaderboard ran at two or more effort levels. We don’t have enough cost data here to say what the extra effort costs.

Lowest effort to highest, light to dark
Claude Fable 5.2
low
100%
Gemini 4 Hydrogen
low
100%
Claude Opus 5
off high
99.1%
Gemini 3.1 Pro
low
98.8%
Gemini 3.8 Flash
low high
98.8%
Gemini 4 Argon
off
98.8%
Gemini 3.7 Flash
low
98.4%
Gemini 3.5 Flash
off high
97.5%
GPT-6.1 Terra
off high
97.5%
GPT-6.2 Sol
low
97.5%
GPT-6.1 Sol
low high
97.2%
GPT-6 Astra
low
96.6%

Source: Official Miðeind Icelandic Case Government leaderboard

ModelEffortHarnessScoreCost per taskOutput tokens
Claude Fable 5.2 high 100%
Gemini 4 Hydrogen high 100%
Claude Fable 5.2 low 99.4%
Gemini 4 Hydrogen low 99.4%
Claude Opus 5 high 99.1%
Gemini 3.1 Pro low 98.8%
Gemini 3.8 Flash high 98.8%
Gemini 4 Argon high 98.8%
Gemini 3.7 Flash high 98.4%
Gemini 4 Carbon low 98.4%
Gemini 4 Neon low 98.4%
Gemini 3.1 Pro high 97.8%
Gemini 3.5 Flash high 97.5%
Claude Opus 5.5 high 97.5%
Gemini 4 Argon off 97.5%
GPT-6.1 Terra high 97.5%
GPT-6.2 Sol low 97.5%
GPT-6.1 Sol high 97.2%
GPT-6.2 Sol high 97.2%
Gemini 3.7 Flash low 96.9%
GPT-6 Astra high 96.6%
Claude Fable 5.1 high 96.3%
GPT-6 Astra low 96.3%
Claude Opus 4.8 high 95.9%
Gemini 3.8 Flash low 95.9%
GPT-5.6 Sol high 95.3%
Claude Sonnet 5.5 high 95.3%
GPT-6 Sol high 94.4%
GPT-6.1 Sol low 94.1%
GPT-6.1 Terra off 93.4%
GPT-5.6 Sol off 90.9%
Claude Sonnet 5 high 89.1%
Claude Opus 4.8 off 88.8%
Gemini 3.5 Flash-Lite high 88.1%
Muse Spark Open high 87.8%
GPT-6 Sol off 87.5%
Inkling 2 Small high 85.6%
Gemini 2.5 Pro Jun 2025 high 84.1%
Gemini 3.5 Flash off 83.1%
Claude Sonnet 5 off 82.2%
Inkling 2 Small minimal 82.2%
Gemini 2.5 Pro Jun 2025 minimal 81.3%
Qwen3.8-2.4T-A95B on 80.3%
Claude Sonnet 5.5 off 80%
GPT-6 Luna high 79.4%
GPT-5.2 high 78.4%
DeepSeek-V4-Pro on 78.4%
DeepSeek V4.1 Flash on 76.9%
Grok 4.7 high 76.9%
Grok 4.7 low 75.3%
GPT-5.6 Terra high 75%
GPT-5.6 Terra off 75%
DeepSeek-V4-Pro off 73.8%
Inkling high 71.3%
GPT-5.6 Luna high 70%
Inkling off 69.4%
Claude Opus 5 off 64.4%
Claude Haiku 5.5 high 64.1%
GPT-5.6 Luna off 63.4%
GPT-6 Luna off 62.8%
GPT-4o Mar 2025 none 61.6%
Gemini 3.5 Flash-Lite off 61.6%
Kimi K3 on 55.6%
GPT-5.2 off 53.8%
DeepSeek V4.1 Flash off 47.2%
Mistral Large 4 high 45%
Muse Spark Open off 44.7%
Inkling-Small high 43.1%
GLM-5.3-Flash on 41.6%
MiniMax-M3 on 33.8%
Mistral Large 4 low 33.4%
Kimi K3 off 26.9%
Mistral Large 3 none 26.6%
Claude Haiku 5.5 off 26.6%
Mistral Small 4 high 25.3%
Gemini 2.5 Flash-Lite Sep 2025 off 24.4%
Qwen 3.8 27B xhigh 21.3%
Gemini 2.5 Flash-Lite Sep 2025 high 18.4%
Mistral Small 4 off 17.5%
Gemma 4 31B IT on 16.6%
Claude Haiku 4.5 off 14.7%
Gemma 4 26B A4B on 14.7%
MiniMax-M3.1-Flash high 10.6%
Qwen 3.8 27B off 9.38%
Claude Haiku 4.5 high 7.81%
Llama 3.1-70B none 7.50%
Mistral Medium 3.5 high 6.25%
Mistral Medium 3.5 off 5.63%
Gemma 4 12B off 5.63%
Gemma 4 26B A4B off 5.31%
Gemma 4 12B on 5.31%
Ripple-65B Preview on 5.31%
Apertus 1.5 8B on 4.69%
MiniMax-M3.1-Flash off 4.38%
Apertus 1.5 70B on 4.06%
GPT-4.1 nano none 3.75%
Gemma 4 E4B on 3.75%
Gemma 4 E4B off 3.44%
GPT-3.5 Turbo Jan 2024 none 2.81%
Llama 3.1-8B none 2.19%
Apertus 8B Instruct none 2.19%
Gemma 4 E2B off 2.19%
Gemma 4 E2B on 2.19%

2 models have held number one on the official leaderboard since its release in September 2026

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

Anthropic
Own leaderboard Passed 85% in Sep 2026, by Claude Opus 5

Source: Official Miðeind Icelandic Case Government leaderboard

Gemini 3.1 Pro on the official leaderboard passed 85% in February 2026

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
On the benchmark’s own leaderboard Best so far
0% 20% 40% 60% 80% 100% 2024 2025 2026 2027 Near the ceiling (85%) Claude Fable 5.2 Gemini 3.1 Pro Gemini 2.5 Pro GPT-4o Llama 3.1-70B GPT-3.5 Turbo 0% 20% 40% 60% 80% 100% 2024 2025 2026 2027 Near the ceiling (85%) Claude Fable 5.2 Gemini 2.5 Pro

Source: Official Miðeind Icelandic Case Government 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 10 checks pass. Every benchmark

Catalog record 2 pass
  • Pass Task access The tasks are private.
  • Pass Confidence Checked against a primary source.
Scores 4 pass 1 note
  • Pass Independent runs Scores from the official leaderboard (62 entries).
  • Pass Models scored 61 models we track, plus 1 other entry on its leaderboard.
  • Pass Newest score The newest is from November 2026, for Gemini 4 Neon.
  • Pass Frontier models 9 of the top 10 have a score, led by Claude Fable 5.2 (#2).
  • Note Headroom Saturated. The best is 100%, past 85% since February 2026.
Leaderboard import 3 pass
  • Pass Importer We read the board on mideind-icelandic-llm-leaderboard.static.hf.space: 103 rows for models we track and 1 other entry 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