VNTL Leaderboard: leaderboard
Translation benchmark where LLMs and translation systems render 256 samples from Japanese visual novels into English while preserving semantic similarity.
Metric: Accuracy (%). Source: huggingface.co. Saturation forecast: Estimated already saturated. 87 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-4o (2024-05-13) | 75.16 | #325 |
| 2 | GPT-4o (2024-08-06) | 74.97 | #336 |
| 3 | Claude 3 Opus | 74.59 | #571 |
| 4 | Claude 3.5 Sonnet (20240620) | 74.4 | #299 |
| 5 | DeepSeek V3 Chat | 74.24 | #447 |
| 6 | Claude 3.5 Sonnet (20241022) | 72.8 | #291 |
| 7 | Nemotron-4 340B Instruct | 72.79 | #641 |
| 8 | GPT-4o Mini (2024-07-18) | 72.23 | #558 |
| 9 | Grok 2 (1212) | 71.6 | #496 |
| 10 | Grok Beta | 71.27 | #487 |
| 11 | DeepSeek V2.5 | 71.14 | #489 |
| 12 | Qwen 2 72B Instruct | 70.2 | #576 |
| 13 | GPT-3.5 Turbo (1106) | 69.98 | #961 |
| 14 | Llama 3.1 70B Instruct | 69.79 | #561 |
| 15 | Llama 3.1 405B Instruct | 69.46 | #457 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=vntl-leaderboard · How It Works · Data refreshed daily, snapshot 2026-10-20.