SEA-HELM (Filipino) - Kalahi (LLM Judge): leaderboard
Kalahi open-ended responses to Filipino cultural situations, graded by an LLM judge. Normalized 0-100 score from AI Singapore's SEA-HELM leaderboard: (score - random baseline) / (maximum - baseline), floored at 0 and averaged over eight runs; higher is better.
Metric: Normalized Score. Source: leaderboard.sea-lion.ai. Saturation forecast: Around February 2027. 58 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemma 4 26B A4B (IT) (Reasoning) | 75.98 | #293 (Gemma 4 26B A4B (IT)) |
| 2 | Qwen 3.5 122B A10B (Reasoning) | 73.82 | #184 (Qwen 3.5 122B A10B) |
| 3 | Qwen 3.6 27B (Reasoning) | 72.37 | #182 (Qwen 3.6 27B) |
| 4 | Gemma 3 27B (IT) | 69.19 | #528 |
| 5 | Qwen 3.6 35B A3B (Reasoning) | 68.91 | #238 (Qwen 3.6 35B A3B) |
| 6 | Qwen 3.5 35B A3B (Reasoning) | 68.58 | #268 (Qwen 3.5 35B A3B) |
| 7 | GPT-OSS-120B (Reasoning) | 67.85 | #346 (GPT-OSS-120B) |
| 8 | Qwen 3.5 27B (Reasoning) | 66.46 | #192 (Qwen 3.5 27B) |
| 9 | Gemma 3 12B (IT) | 62.69 | #672 |
| 10 | Llama-SEA-LION-v3-70B-IT | 61.37 | #407 |
| 11 | Llama 3.3 70B Instruct | 57.27 | #541 |
| 12 | Qwen 3 VL 32B Instruct | 55.21 | #290 |
| 13 | Mistral Small 4 | 54.8 | #522 |
| 14 | Qwen 3 32B (Thinking) | 54.5 | #449 (Qwen 3 32B) |
| 15 | Qwen 3 14B (Reasoning) | 53.21 | #542 (Qwen 3 14B) |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=sea-helm-filipino-kalahi-llm-judge · How It Works · Data refreshed daily, snapshot 2026-11-05.