kth8/Qwen3.6-35B-A3B-SuperGPQA-benchmark

Benchmark of Qwen/Qwen3.6-35B-A3B against m-a-p/SuperGPQA dataset. Accuracy: 64.8% with Python tool. Metric Value Correct 648 Incorrect 337 Errors 15 Total samples 1000 Python tool calls 1473 Total completion tokens 4,837,137 Raw stats: { "accuracy": 0.648, "correct": 648, "incorrect": 337, "error": 15, "total": 1000, "python_tool_calls": 1473, "completion_tokens": 4837137 }