FilBench - Reading Comprehension: leaderboard
From the FilBench leaderboard (Miranda et al., 2025), which evaluates LLMs on Filipino, Tagalog and Cebuano with a lighteval fork. Reading Comprehension combines Belebele (Cebuano and Filipino), NewsPH NLI and Cebuano readability classification, 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 | GPT-4o (2024-08-06) | 80.12 | #343 |
| 2 | Qwen 2.5 72B Instruct | 75.62 | #469 |
| 3 | Llama 4 Maverick Instruct FP8 | 72.99 | #424 |
| 4 | Llama 4 Scout Instruct | 70.86 | #547 |
| 5 | Qwen 2.5 32B Instruct | 70.59 | #509 |
| 6 | Qwen 3 8B | 67.3 | #678 |
| 7 | Qwen 3 14B | 66.11 | #542 |
| 8 | Qwen 3 32B | 65.48 | #449 |
| 9 | Mixtral 8x22B Instruct | 64.78 | #738 |
| 10 | Sailor2-20B-Chat | 63.03 | #589 |
| 11 | SeaLLMs-v3-7B-Chat | 62.47 | #897 |
| 12 | Ministral-8B-Instruct-2410 | 62.33 | #1081 |
| 13 | Mixtral 8x7B Instruct | 60.95 | #1085 |
| 14 | Qwen 2.5 7B Instruct | 60.47 | #863 |
| 15 | Qwen 2.5 14B Instruct | 59.95 | #654 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=filbench-reading-comprehension · How It Works · Data refreshed daily, snapshot 2026-11-05.