GPT-6 Astra beats Fable 5.1 and Gemini 3.8 Flash in 3D library reconstruction test
ZhitingHu · x · 2026-09-05
EnactraAI stress-tested GPT-6 Astra against GPT-5.6 Sol, Fable 5.1, and Gemini 3.8 Flash on agentic 3D reconstruction, tasking each with rebuilding the Geisel Library from photos alone.
- Astra led on overall structure, geometry, and fine details, delivering notably higher visual fidelity than the other flagships
- This is among the first third-party agentic 3D benchmarks released during GPT-6 Astra's launch week
- The comparison is presented as a side-by-side video highlighting geometry, structure, and rendering differences
Related event: Flagship Models Tested on Photo-Based 3D Reconstruction, GPT-6 Astra Leads(2 posts)→
More from Models
- Users flag: Claude locks you out of your own Projects after subscription cancellation — BLUECOW009 · 2026-09-05
- Dev says GPT-6 Astra found 4.5x to 176x speedups in his codebase in 5 minutes — charliermarsh · 2026-09-05
- Dev calls Astra's computer use "superhuman, not human level" — intellectronica · 2026-09-05
- Codex's New Voice Mode Detects Sniffles, Coughs, and Gulp Sounds — craigsdennis · 2026-09-05
- Google shipped four Gemini Flash models in 106 days, flagship 3.5 Pro still AWOL — fortune · 2026-09-05
- GPT Astra beats Claude Fable 5.1 with 33% fewer tokens and 33% lower cost — rohanpaul_ai · 2026-09-05