SEA-HELM (English) - MuSR: leaderboard
MuSR multistep soft-reasoning problems (murder mysteries, object placement, team allocation) run by SEA-HELM. 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 | Qwen 3.6 27B (Reasoning) | 81.06 | #182 (Qwen 3.6 27B) |
| 2 | Qwen 3.6 35B A3B (Reasoning) | 79.32 | #238 (Qwen 3.6 35B A3B) |
| 3 | Gemma 4 26B A4B (IT) (Reasoning) | 79.25 | #293 (Gemma 4 26B A4B (IT)) |
| 4 | Qwen 3.5 27B (Reasoning) | 77.73 | #192 (Qwen 3.5 27B) |
| 5 | GPT-OSS-120B (Reasoning) | 77.64 | #346 (GPT-OSS-120B) |
| 6 | Qwen 3.5 122B A10B (Reasoning) | 76.61 | #184 (Qwen 3.5 122B A10B) |
| 7 | Qwen 3.5 35B A3B (Reasoning) | 75.94 | #268 (Qwen 3.5 35B A3B) |
| 8 | Qwen 3.5 9B (Reasoning) | 75.6 | #391 (Qwen 3.5 9B) |
| 9 | Qwen 3 32B (Thinking) | 75.2 | #449 (Qwen 3 32B) |
| 10 | Qwen 3 14B (Reasoning) | 73.59 | #542 (Qwen 3 14B) |
| 11 | Qwen 3.5 4B (Reasoning) | 73.11 | #514 (Qwen 3.5 4B) |
| 12 | Qwen 3 8B (Thinking) | 71.93 | #678 (Qwen 3 8B) |
| 13 | Qwen 3 VL 32B Instruct | 71.26 | #290 |
| 14 | GLM-4.7 Flash (Reasoning) | 67.8 | #543 (GLM-4.7 Flash) |
| 15 | Llama-SEA-LION-v3-70B-IT | 65.23 | #407 |
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-english-musr · How It Works · Data refreshed daily, snapshot 2026-11-05.