HELM SeaHELM - XCOPA (th): leaderboard
HELM SeaHELM run of XCOPA, Thai causal commonsense reasoning (exact match).
Metric: EM. Source: crfm.stanford.edu. Saturation forecast: Estimated already saturated. 21 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude 3.7 Sonnet (20250219) | 97.8 | #201 |
| 2 | Qwen 2.5 72B Instruct | 97.4 | #454 |
| 3 | GPT-4o (2024-11-20) | 97.2 | #381 |
| 4 | Llama 3.1 70B Instruct | 96.2 | #561 |
| 5 | Llama 3.3 70B Instruct | 96 | #534 |
| 6 | Llama 3.2 90B Vision Instruct | 96 | #622 |
| 7 | Llama 3.1 405B Instruct | 95.8 | #457 |
| 8 | Gemini 2.0 Flash | 95.6 | #342 |
| 9 | DeepSeek V3 | 95.4 | #321 |
| 10 | Gemini 2.0 Flash Lite | 95.2 | #450 |
| 11 | Qwen 2 72B Instruct | 94.4 | #576 |
| 12 | GPT-4o Mini (2024-07-18) | 93.6 | #558 |
| 13 | Llama 4 Maverick Instruct FP8 | 86.4 | #414 |
| 14 | Claude 3.5 Haiku (20241022) | 86.2 | #588 |
| 15 | Llama 3.1 8B CPT SEA LIONv3 Instruct | 83.4 | #810 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=helm-seahelm-xcopa-th · How It Works · Data refreshed daily, snapshot 2026-10-20.