GPT-6.1 Terra beats Sonnet 5.5 on same Blender task: 10 min/$0.31 vs 1.5 hr/$15.20
CtrlAltDwayne · x · 2026-10-27
A side-by-side test had GPT-6.1 Terra (High) and Claude Sonnet 5.5 (High) build the same "red supercar through Miami at night" scene from scratch in Blender. Terra finished in 10 minutes at $0.31; Sonnet 5.5 took 1h26m at $15.20 — roughly a 50x gap. The reposter admits initial skepticism about OpenAI's benchmarks but concedes the model is delivering after seeing community results.
More from Models
- DeepSeek site shows new v4-flash fingerprints as community cracks checkpoint names — teortaxesTex · 2026-10-27
- User claims Kimi K3.1 Pro High generated an LLM explainer video; model name unverified — iamfakhrealam · 2026-10-27
- Anyone can get an abliterated model with safety training stripped out for pennies — EXM7777 · 2026-10-27
- Ask an LLM for 10 bold 2030 AI predictions — many already happened — FlorianGallwitz · 2026-10-27
- Prefilling probe shows curl trigger: 98% completions write a URL once '-o /tmp/' is appended — rickasaurus · 2026-10-27
- Unsloth details running DeepSeek-V4.1-Flash locally: 1-bit GGUF fits 128GB RAM with 1M context — gnukeith · 2026-10-27