FilBench - Cultural Knowledge: leaderboard
From the FilBench leaderboard (Miranda et al., 2025), which evaluates LLMs on Filipino, Tagalog and Cebuano with a lighteval fork. Cultural Knowledge combines Global-MMLU (Tagalog), INCLUDE, KALAHI and StingrayBench semantic appropriateness, weighted by instance count; accuracy 0-100, higher is better. Runs missing one of these tasks are omitted.
Metric: Category Score (%). Source: huggingface.co. Saturation forecast: Estimated already saturated. 38 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Llama-SEA-LION-v3-70B-IT | 76.78 | #407 |
| 2 | Llama 4 Maverick Instruct FP8 | 76.75 | #424 |
| 3 | Llama 4 Scout Instruct | 74.31 | #547 |
| 4 | GPT-4o (2024-08-06) | 73.29 | #343 |
| 5 | Qwen 2.5 72B Instruct | 73.11 | #469 |
| 6 | Llama 3.1 70B Instruct | 72.16 | #567 |
| 7 | Gemma 3 27B (IT) | 71.41 | #528 |
| 8 | Qwen 3 32B | 71.29 | #449 |
| 9 | c4ai-command-a-03-2025 | 69.95 | #468 |
| 10 | Gemma 2 27B (IT) | 68.76 | #652 |
| 11 | Gemma 3 12B (IT) | 67.34 | #672 |
| 12 | Qwen 2.5 32B Instruct | 66.83 | #509 |
| 13 | Sailor2-20B-Chat | 66.43 | #589 |
| 14 | gemma2-9B-cpt-sea-lionv3-instruct | 64.44 | #717 |
| 15 | Gemma 2 9B (IT) | 63.69 | #795 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=filbench-cultural-knowledge · How It Works · Data refreshed daily, snapshot 2026-11-05.