ChatGPT 5.6 Sol Still Hallucinates: 3 Real-World Failure Cases
kaljakin · reddit · 2026-08-26
A Reddit user critiques ChatGPT 5.6 Sol, sharing three failure cases despite its long "thinking" time:
- Flawed Logic: Asked for a small car for narrow streets, it recommended a larger Hyundai i30 instead of a MINI Cooper. When challenged, it bizarrely claimed that 3-door cars have longer doors than 5-door ones based on a flawed geometric argument.
- Search Failure: It recommended a cave ("Grgosova špilja") for a return trip, failing to notice clear text and image info on Google Maps stating it was closed for the holidays, despite being asked about opening hours.
- Data Task: The user also attempted to use the model to script a market-price-based anomaly detector for supplier price lists.
More from Models
- Grok Voice 3.0 reportedly powers Tesla in-car assistant with millions of daily conversations — XFreeze · 2026-10-15
- Gemini 4 Argon coding style: neat one-line comments, but does the code actually work? — zacharynado · 2026-10-15
- Why ChatGPT beats the better Claude with mainstream users: being first known — aitrendz_xyz · 2026-10-15
- 'You need high pain tolerance to review Gemini 4,' researcher quips amid error screens — gleech · 2026-10-15
- MiniCPM-V 4.7 runs a 35B MoE multimodal model on phones with just 3GB memory — nevrekaraishwa2 · 2026-10-15
- Gemini 4 Argon wrote 5,000+ lines of pure Python to build its own ray-tracing 3D engine — aitrendz_xyz · 2026-10-15