Benchmark Almanac

Kinayat

Claude 3.5 Sonnet 95.3% leads since October 2025
Not in model cards

Models interpret Egyptian Arabic idioms, select appropriate contexts and identify connotations. The evaluation includes separate generation and multiple-choice tasks.

A later study by three original authors evaluated four model families on 150 idioms. It measured accuracy on two possible explanations.

Released
October 2025
Kind
Reasoning in many languages
Level
Everyday
Scored
Checked automatically
Answers
Multiple choice
Tasks
All public
Measures
Capability
Language
Arabic
In and out
Text → Text
Interaction
Single turn
Refresh
One fixed set
Run by
Anyone
Strongest model tested
Claude Sonnet 4 #113 on the capability index
Last model added
September 2025
Scores refreshed
November 21st, 2026
22
Entries in its paper, 8 models we track

3 models share the top score of 95.3% among models released by September 2025 3 models share the top score of 95.3% among models released by September 2025

Best score per entry from the benchmark's paper, read on November 21st, 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 September 2025, 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 95.3%
  2. 2 GPT-4o Mar 2025 95.3%
  3. 3 Claude Sonnet 4 95.3%
  4. 4 Gemini 2.5 Flash-Lite Sep 2025 93.3%
  5. 5 Llama-3.1-70B-Instruct 88.7%
  6. 6 ALLaM-7B-Instruct-preview 85.3%
  7. 7 GPT-4o mini 85.3%
  8. 8 Aya Expanse 32B 84.7%
  9. 9 Qwen2.5-14B-Instruct 84%
  10. 10 Gemini 1.5 Flash Sep 2024 84%
  11. 11 Qwen2.5-32B-Instruct 82%
  12. 12 Gemma-2-27b-it 81.3%
  13. 13 Gemma-2-9B-it 78.7%
  14. 14 Jais-family-13b-chat 78.7%
  15. 15 Fanar-1-9B-Instruct 77.3%
  16. 16 Qwen2.5-7B-Instruct 77.3%
  17. 17 Jais-family-6p7b-chat 70.7%
  18. 18 Aya Expanse 8B 65.3%
  19. 19 Llama-3.1-8B-Instruct 60%
  20. 20 SILMA-9B-Instruct-v1.0 59.3%
  21. 21 AceGPT-v2-8B-Chat 56.7%
  22. 22 Mistral-7B-Instruct-v0.3 48.7%

Source: The Kinayat paper

#EntryScoreSettingDateSource
1 Claude 3.5 Sonnet 95.3% Jun 2024 Official leaderboard
2 GPT-4o Mar 2025 95.3% Mar 2025 Official leaderboard
3 Claude Sonnet 4 95.3% May 2025 Official leaderboard
4 Gemini 2.5 Flash-Lite Sep 2025 93.3% Sep 2025 Official leaderboard
5 Llama-3.1-70B-Instruct 88.7% Official leaderboard
6 ALLaM-7B-Instruct-preview 85.3% Official leaderboard
7 GPT-4o mini 85.3% Jul 2024 Official leaderboard
8 Aya Expanse 32B 84.7% Oct 2024 Official leaderboard
9 Qwen2.5-14B-Instruct 84% Official leaderboard
10 Gemini 1.5 Flash Sep 2024 84% Aug 2024 Official leaderboard
11 Qwen2.5-32B-Instruct 82% Official leaderboard
12 Gemma-2-27b-it 81.3% Official leaderboard
13 Gemma-2-9B-it 78.7% Official leaderboard
14 Jais-family-13b-chat 78.7% Official leaderboard
15 Fanar-1-9B-Instruct 77.3% Official leaderboard
16 Qwen2.5-7B-Instruct 77.3% Official leaderboard
17 Jais-family-6p7b-chat 70.7% Official leaderboard
18 Aya Expanse 8B 65.3% Oct 2024 Official leaderboard
19 Llama-3.1-8B-Instruct 60% Official leaderboard
20 SILMA-9B-Instruct-v1.0 59.3% Official leaderboard
21 AceGPT-v2-8B-Chat 56.7% Official leaderboard
22 Mistral-7B-Instruct-v0.3 48.7% Official leaderboard
#ModelScoreSettingDateSource
1 Claude 3.5 Sonnet 95.3% Jun 2024 Official leaderboard
2 GPT-4o Mar 2025 95.3% Mar 2025 Official leaderboard
3 Claude Sonnet 4 95.3% May 2025 Official leaderboard
4 Gemini 2.5 Flash-Lite Sep 2025 93.3% Sep 2025 Official leaderboard
5 GPT-4o mini 85.3% Jul 2024 Official leaderboard
6 Aya Expanse 32B 84.7% Oct 2024 Official leaderboard
7 Gemini 1.5 Flash Sep 2024 84% Aug 2024 Official leaderboard
8 Aya Expanse 8B 65.3% Oct 2024 Official leaderboard

Rankings by part

We read 2 parts from the benchmark’s paper: Pragmatic Use and Contextual Understanding. Each ranks models on one part of the test. Claude 3.5 Sonnet leads Pragmatic Use and Gemini 2.5 Flash-Lite leads Contextual Understanding. Arrows show how far a model moved from its overall rank among the models we track.

Pragmatic Use

8 models we track have a score here.

  1. 1 Claude 3.5 Sonnet 85.3%
  2. 2 GPT-4o Mar 2025 84%
  3. 3 Claude Sonnet 4 83.3%
  4. 4 Gemini 2.5 Flash-Lite Sep 2025 78%
  5. 5 Aya Expanse 32B ↑1 75.3%
  6. 6 Gemini 1.5 Flash Sep 2024 ↑1 65.3%
  7. 7 GPT-4o mini ↓2 64.7%
  8. 8 Aya Expanse 8B 56%

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 Oct 2025, by Claude 3.5 Sonnet

Source: The Kinayat paper

Claude 3.5 Sonnet on the official leaderboard passed 85% in June 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
0% 20% 40% 60% 80% 100% 2024 2025 Near the ceiling (85%) Claude 3.5 Sonnet 0% 20% 40% 60% 80% 100% 2024 2025 Near the ceiling (85%) Claude 3.5 Sonnet

Source: Official Kinayat 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 (22 entries).
  • Pass Models scored 8 models we track, plus 14 other entries on its leaderboard.
  • Note Newest score The newest came 14 months ago, in September 2025, for Gemini 2.5 Flash-Lite. 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 Sonnet 4 (#113).
  • Note Headroom Saturated. The best is 95.3%, past 85% since June 2024.
Leaderboard import 3 pass
  • Pass Importer We read the board on aclanthology.org: 8 rows for models we track and 14 other entries at the last read.
  • Pass Last read November 21st, 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