I ran extensive tests on this and variations on multiple models. Most models interpret 50 m as a short distance and struggle with spatial reasoning. Only Gemini and Grok correctly inferred that you would need to bring your car to get it washed in their thought stream, and incorporated that into the final answer. GPT-5.2 and Kimi K2.5 and even Opus 4.6 failed in my tests - https://x.com/sathish316/status/2023087797654208896?s=46
What surprised me was how introducing a simple, seemingly unrelated context - such as comparing a 500 m distance to the car wash to a 1 km workout - confused nearly all the models. Only Gemini Pro passed my second test after I added this extra irrelevant context - https://x.com/sathish316/status/2023073792537538797?s=46
Most real-world problems are messy and won’t have the exact clean context that these models are expecting. I’m not sure how the major AI labs assume most real-world problems are simpler than the constraints exposed by this example like prerequisites, ordering, and contextual reasoning, which are already posing challenges to these bigger models.
