Benchmark Almanac

LingOly-TOO

GPT-5 46.7% leads since August 2025
Not in model cards

Models solve Linguistics Olympiad problems whose spelling patterns have been systematically changed while preserving the solution logic. Answers receive exact-match credit with no partial credit. The evaluation compares performance on original and obfuscated problems.

The May 2026 paper revision reports 1,203 questions and 6,995 subquestions. Obfuscations alter the task inputs to reduce memorization shortcuts. The linked Space was sleeping during this review, so the paper and dataset supply the verified evidence.

Released
March 2025
Kind
Reasoning in many languages
Level
Competition
Scored
Checked automatically
Answers
Short answer
Tasks
All public
Measures
Capability
Language
Many languages
In and out
Text → Text
Interaction
Single turn
Items from
Exams and contests
Refresh
One fixed set
Run by
Anyone
Based on
LingOly
In views
Many languages
Strongest model tested
GPT-5 #71 on the capability index
Last model added
September 2025
Scores refreshed
November 19th, 2026

In the authors’ words

Frontier language models demonstrate increasing ability at solving reasoning problems, but their performance is often inflated by circumventing reasoning and instead relying on their expanding knowledge and memorisation capacity. We introduce LINGOLY-TOO, a challenging reasoning benchmark of 1,203 questions and a total of 6,995 sub-questions that counters these shortcuts by applying expert-designed obfuscations to Linguistics Olympiad problems. These obfuscations preserve the underlying solution logic while reducing the likelihood problems are solvable with via knowledge or memorisation. Our experiments show that models exploit shortcuts on the original question as performance markedly drop upon obfuscation. Even the best reasoning models remain highly sensitive, with scores dropping from around 0.59 on original problems to 0.48 after obfuscation. LINGOLY-TOO disentangles reasoning from knowledge, offering a clearer measure of true reasoning capabilities.

Abstract of LINGOLY-TOO: Disentangling Reasoning from Knowledge with Templatised Orthographic Obfuscation (arXiv:2503.02972, CC0)
15
Entries on its own leaderboard, 14 models we track

GPT-5 leads with 46.7% among models released by September 2025 GPT-5 leads with 46.7% among models released by September 2025

Best score per entry from the benchmark's own leaderboard, read on November 19th, 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. 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 GPT-5 46.7%
  2. 2 Claude Opus 4.1 45.8%
  3. 3 Claude 3.7 Sonnet 42.9%
  4. 4 Gemini 2.5 Pro Jun 2025 42.4%
  5. 5 DeepSeek-V3.1-Terminus 42.2%
  6. 6 o1-preview 32.2%
  7. 7 o3-mini high 30.6%
  8. 8 Claude 3.5 Sonnet 28.1%
  9. 9 DeepSeek-R1 May 2025 26.5%
  10. 10 GPT-4.5 25.4%
  11. 11 Gemini 1.5 Pro Sep 2024 20.5%
  12. 12 GPT-4o Mar 2025 15.6%
  13. 13 Phi-4 11%
  14. 14 Llama 3.3 70B Instruct 8.21%
  15. 15 Aya 23 35B 5.71%

Source: Official LingOly-TOO leaderboard

#EntryScoreSettingDateSource
1 GPT-5 46.7% Aug 2025 Official leaderboard
2 Claude Opus 4.1 45.8% Aug 2025 Official leaderboard
3 Claude 3.7 Sonnet 42.9% Feb 2025 Official leaderboard
4 Gemini 2.5 Pro Jun 2025 42.4% Jun 2025 Official leaderboard
5 DeepSeek-V3.1-Terminus 42.2% Sep 2025 Official leaderboard
6 o1-preview 32.2% Sep 2024 Official leaderboard
7 o3-mini 30.6% high Jan 2025 Official leaderboard
8 Claude 3.5 Sonnet 28.1% Jun 2024 Official leaderboard
9 DeepSeek-R1 May 2025 26.5% May 2025 Official leaderboard
10 GPT-4.5 25.4% Feb 2025 Official leaderboard
11 Gemini 1.5 Pro Sep 2024 20.5% Sep 2024 Official leaderboard
12 GPT-4o Mar 2025 15.6% Mar 2025 Official leaderboard
13 Phi-4 11% Dec 2024 Official leaderboard
14 Llama 3.3 70B Instruct 8.21% Dec 2024 Official leaderboard
15 Aya 23 35B 5.71% Official leaderboard
#ModelScoreSettingDateSource
1 GPT-5 46.7% Aug 2025 Official leaderboard
2 Claude Opus 4.1 45.8% Aug 2025 Official leaderboard
3 Claude 3.7 Sonnet 42.9% Feb 2025 Official leaderboard
4 Gemini 2.5 Pro Jun 2025 42.4% Jun 2025 Official leaderboard
5 DeepSeek-V3.1-Terminus 42.2% Sep 2025 Official leaderboard
6 o1-preview 32.2% Sep 2024 Official leaderboard
7 o3-mini 30.6% high Jan 2025 Official leaderboard
8 Claude 3.5 Sonnet 28.1% Jun 2024 Official leaderboard
9 DeepSeek-R1 May 2025 26.5% May 2025 Official leaderboard
10 GPT-4.5 25.4% Feb 2025 Official leaderboard
11 Gemini 1.5 Pro Sep 2024 20.5% Sep 2024 Official leaderboard
12 GPT-4o Mar 2025 15.6% Mar 2025 Official leaderboard
13 Phi-4 11% Dec 2024 Official leaderboard
14 Llama 3.3 70B Instruct 8.21% Dec 2024 Official leaderboard

3 models have held number one on the official leaderboard since its release in March 2025

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

Anthropic OpenAI
Own leaderboard Not tested since Sep 2025
GPT-5 46.7%

Source: Official LingOly-TOO 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 the benchmark’s own leaderboard Best so far
0% 20% 40% 60% 80% 100% 2024 2025 Near the ceiling (85%) GPT-5 Claude 3.7 Sonnet o1-preview 0% 20% 40% 60% 80% 100% 2024 2025 Near the ceiling (85%) GPT-5 o1-preview

Source: Official LingOly-TOO 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 14 months ago, in September 2025, for DeepSeek-V3.1-Terminus. 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 (15 entries).
  • Pass Models scored 14 models we track, plus 1 other entry on its leaderboard.
  • Warning Newest score The newest came 14 months ago, in September 2025, for DeepSeek-V3.1-Terminus.
  • Note Frontier models None of the top 10 has a score. The strongest tested is GPT-5 (#71).
  • Pass Headroom The best is 46.7%, under 85%, so it still separates models.
Leaderboard import 3 pass
  • Pass Importer We read the board on huggingface.co: 15 rows for models we track and 1 other entry at the last read.
  • Pass Last read November 19th, 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