Reading up on gpt-oss-120B

Jev, a constrained decision model, matched Grok's 89.6% accuracy on 48 game questions while cutting latency 84% and cost 88.5% by using a selective Gemini fallback.

Alibaba's Qwen3.5-9B outperforms OpenAI's 120B-parameter gpt-oss on benchmarks while being 13 times smaller and capable of running on standard laptops.