Pixel art benchmark: Gemini 4 Neon wins on quality but costs 13x more and 4x slower than GPT-6.2 Sol
Angaisb_ · x · 2026-11-22
A hands-on 32×32 pixel-art comparison (high reasoning, one object per request, unrestricted palette) finds Gemini 4 Neon produces the best images at a steep cost.
Totals across six objects (kingfisher, rotary phone, peeled lemon, pocket watch, fly agaric, porcelain teapot):
- Gemini 4 Neon (no colors instructions): $3.026 total, 307.4 s average per image; priciest single image $0.667, slowest 381.2 s
- GPT-6.2 Sol (with colors instructions): $0.237 total, 65 s average — fastest and cheapest
- GPT-6.2 Astra (with colors instructions): $1.030 total, 95.9 s average
Takeaway: Neon clearly leads on pixel-art detail and texture, but costs 13x more and runs 3-5x slower than Sol, making Sol/Astra the rational default unless quality is paramount.
More from Models
- Claude concedes a scalar-bound mistake after reading the author's blog post, keeping its curvature arguments — bremen79 · 2026-11-22
- Uncensored Inkling 2 Mini v2: self-distillation lifts useful answers from 52% to 79% with open pipeline — Turbulent-Sky5396 · 2026-11-22
- Looped transformers explained: Inkling 2 Mini loops 12 middle layers 3x for deeper reasoning without more weights — Prompt Engineering · 2026-11-22
- GPT-5.4 with open-source KYNEX harness outscored standalone GPT-6 Astra on ExtractBench: 92.97 vs 91.9 F1 — the-random-guy-2002 · 2026-11-22
- Researcher claims Gemini Neon helped him land a potential Nature-worthy physics idea — xennygrimmato_ · 2026-11-22
- Unnecessary Hyphens and Semicolons Are the New En-Dash in Claude Writing — bttyeo · 2026-11-22