OpenAI's GPT-OSS models benchmarks worse than DeepSeek R1 and Qwen3 235B xcancel.com 8 points by pu_pe a day ago
[dead]