IOL-AI Challenge 2026: leaderboard
The five unseen problems of the International Linguistics Olympiad 2026 Individual Contest, where the model works out an unfamiliar language's grammar and vocabulary from a small set of worked examples. Automatic score after the paper's answer parsing: the geometric mean of chrF and exact match. Higher is better. Source: arxiv.org/abs/2608.18011.
Metric: Geometric Mean of chrF and Exact Match (%; IOL 2026 Individual Contest, automatic). Source: arxiv.org. Saturation forecast: Around February 2027. 18 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Opus 4.8 | 75.1 | #57 |
| 2 | Gemini 3.6 Flash | 56.4 | #74 |
| 3 | GPT-5.6 Sol | 52.2 | #29 |
| 4 | Kimi K2.6 | 45.2 | #176 |
| 5 | GLM-5.2 | 45 | #138 |
| 6 | DeepSeek V4 Pro | 41 | #188 |
| 7 | Qwen 3.8 27B | 39.8 | #206 |
| 8 | Qwen 3.7 Max | 39.2 | #143 |
| 9 | Muse Glimmer 30B (High) | 38.4 | #381 (Muse Glimmer 30B) |
| 10 | Inkling | 34.1 | #271 |
| 11 | Nemotron 3 Ultra | 28.9 | #427 |
| 12 | Qwen 3.6 27B (Thinking) | 26.7 | #345 (Qwen 3.6 27B) |
| 13 | Gemma 4 31B | 23.2 | #392 |
| 14 | Command A+ | 19.4 | #940 |
| 15 | Qwen 3.6 27B (Non-reasoning) | 19.4 | #345 (Qwen 3.6 27B) |
No result here: #4 Claude Opus 5.5, #5 GPT-6 Astra, #9 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=iol-ai-challenge-2026 · How It Works · Data refreshed daily, snapshot 2026-11-08.