Konkur 1404 - Overall: leaderboard
Konkur 1404 — multiple-choice questions from the 1404 (2025) edition of the Konkur, Iran's national university entrance exam, posed in Persian; overall text-only accuracy (%) across all exam streams.
Metric: Accuracy (%, text-only). Source: huggingface.co. Saturation forecast: Around June 2028. 23 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-OSS-120B | 53.6 | #341 |
| 2 | gpt-oss-safeguard-20B | 51 | #314 |
| 3 | Qwen 3 VL 235B A22B Instruct | 50.54 | #273 |
| 4 | Kimi K2 0905 | 47.17 | #288 |
| 5 | Qwen 3 Next 80B A3B Instruct | 45.79 | #368 |
| 6 | Devstral Small 2 | 43.8 | #495 |
| 7 | Qwen 3 VL 30B A3B Instruct | 43.64 | #384 |
| 8 | DeepSeek V3.2 | 42 | #205 |
| 9 | Gemma 3 27B (IT) | 41.81 | #523 |
| 10 | Llama 3.1 70B Instruct | 40.74 | #561 |
| 11 | Ministral 3 14B | 39.82 | #650 |
| 12 | Llama 4 Maverick | 39.51 | #464 |
| 13 | MiMo-V2-Flash | 37.37 | #372 |
| 14 | Mixtral 8x22B Instruct | 36.45 | #728 |
| 15 | Gemma 3n E4B (IT) | 34 | #749 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=konkur-1404-overall · How It Works · Data refreshed daily, snapshot 2026-10-20.