GPT-5.6 Sol Ultra Lost to GPT-5.5 in Physics Tests
Testing of the GPT-5.6 Sol Ultra and GPT-5.5 models revealed that the new model underperformed in physics simulation despite its higher cost. The results indicate that the previous version remains more effective in task execution.
GPT-5.6 Sol Ultra Lost to GPT-5.5 in Physics Tests
Testers provided four models with the same prompt: create three autonomous scenes on HTML5 Canvas with realistic physics simulation.
Prompts:
- A monster truck does a backflip and lands on a parked car.
- A stunt car jumps over six buses and crashes into a brick wall.
- A train derails on a broken bridge and falls into the water.
Results:
- GPT-5.6 Sol Ultra: 32.9k tokens, $0.33
- Opus 4.8: 9.2k tokens, $0.24
- GPT-5.5: 12.4k tokens, $0.11
- Grok 4.5: 7.0k tokens, $0.08
Sol Ultra draws almost as well as GPT-5.5, but with greater detail. This is especially noticeable in the bus-jumping scene, where the results of the two models look almost identical. In other tests, Sol Ultra performed worse than its predecessor, and the quality of physics simulation noticeably decreased. In their view, GPT-5.5 confidently wins in the scenes with the monster truck flip and the train crash.
They concluded that Sol Ultra is essentially GPT-5.5 with weaker physics but a prettier picture, and it is three times more expensive. 🤣
The recently released Grok 4.5 failed two out of three tests and only performed adequately in the train scene.
Why it matters
AnalysisThis is important as comparing AI models helps understand their strengths and weaknesses, influencing technology choices for development. The testing results can impact companies' decisions regarding the adoption of new solutions in their projects.
Discuss in community
Share your questions and insights with developers
Related in Ecosystem
Practical guides, core terminology, and agent skills
JSON — Structured Data for Programs, APIs, and AI Agents
A complete hands-on guide to JSON: syntax specifications, serialization via parse and stringify, REST API consumption, JSON Schema validation, and structured outputs for LLMs.
DOCX and XLSX Formats: When AI Needs to Deliver Production-Ready Documents
A comprehensive hands-on guide to generating and processing DOCX documents and XLSX workbooks with AI: comparing against Markdown and CSV, prompt engineering, formula design, and code automation.
Binance publishes investigation into $6.3 billion theft
Binance’s investigation links the theft to a modified Homebrew download, a fake interview and abuse of an internal AI tool.
Reddit user explains why Claude Code weekly limits are burning almost 4x faster
A Reddit user compared token usage across two weeks and found that each 1% of Claude Code's weekly limit now yields far fewer output tokens, pointing to subagents, reasoning time, and UI refresh requests as likely causes.