Photo-to-Blender benchmark: GPT-6 Astra swept every photo, GPT-6.1 Sol scored 61 for 36 cents
smith2008 · reddit · 2026-10-02
The author's OpenAI-side results from a photo-to-Blender agent benchmark (same loop, 20-min/$4/60-request caps, deterministic 0-100 scorer):
| Model | Score | Cost | Time |
|---|---|---|---|
| GPT-6 Astra | 66 | $3.91 | 18 min |
| GPT-6.1 Sol | 61 | $0.36 | 18 min |
| GPT-5.6 Sol | 49 | $0.57 | 8 min |
| GPT-5.6 Terra | 44 | $0.33 | 8 min |
Astra placed first on all three photos and was stopped by the cost cap every time — first render at 4 minutes, refining until the gateway refused requests at $3.90. GPT-6.1 Sol worked the full 18 minutes, finishing 5 points behind for under a tenth of the price. GPT-5.6 models declared completion after 8 minutes, leaving most budget unspent; their scenes are simpler, not broken.
Caveats: one run per model per photo, and the brief was shaped around Astra.
Related event: GPT-6 Astra Sweeps 14-Model 3D Reconstruction Benchmark(3 posts)→
More from coding & agent
- Mirelo brings AI sound effects generation to AWS's Kiro agentic IDE — ordax · 2026-10-02
- Using Codex 8 hours a day and still can't burn through the limits — honkballs · 2026-10-02
- Free 18k-star GitHub course on Harness Engineering: 14 lectures + 8 hands-on projects — Hesamation · 2026-10-02
- Manifesto: open-source project makes UI and agents share the same app-owned actions — TraditionalListen994 · 2026-10-02
- Debian kernel alert teems with void* bugs: who (or what) is auditing the Linux kernel? — mircomusolesi · 2026-10-02
- Google's A2A protocol sees hype but little production use yet — Diarnstinc_Crew_6510 · 2026-10-02