kth8/Qwen3.5-9B-SuperGPQA-benchmark

--- license: apache-2.0 language: - en base_model: Qwen/Qwen3.5-9B datasets: - m-a-p/SuperGPQA --- Benchmark of [Qwen/Qwen3.5-9B](https://huggingface.co/Qwen/Qwen3.5-9B) against [m-a-p/SuperGPQA](https://huggingface.co/datasets/m-a-p/SuperGPQA) dataset. Accuracy: 50.3% with Python tool. | Metric | Value | |----------------------|---------------| | **Correct** | 503 | | **Incorrect** | 494 | | **Errors** | 3 | | **Total samples** | 1000 | | **Python tool calls**| 1454 | | **Total completion tokens** | 2,958,569 | Raw stats: ```json { "accuracy": 0.503, "correct": 503, "incorrect": 494, "error": 3, "total": 1000, "python_tool_calls": 1454, "completion_tokens": 2958569 } ```