I was very excited to try Gemini 4 Argon this morning. It definitely makes bette
By Drmikerand · October 15, 2026 · Curated by George's Blog
I was very excited to try Gemini 4 Argon this morning. It definitely makes better planograms. I can't believe that the general release of Gemini 4 Argon wasn't the headline item on the Ten O'clock News yesterday. Not even a mention!
The latest development version of Spanner The Space Planner is now AI enabled and I gave Gemini 4 Argon and Gemini 3.8 Flash exactly the same prompt "create a new planogram with best sellers". I had a document loaded for each with some fake sales data and a Tesco fixture template.
Argon took just under 4 minutes to produce a planogram and cost $0.71.
Flash took just over 5 minutes but only cost $0.47 in tokens/compute.
But the Argon planogram is definitely better, here is what Claude had to say:
Argon is significantly more space-efficient — it squeezes in a 6th shelf by tightening inter-shelf gaps, resulting in more product variety and substantially higher stock capacity.
• Argon has tight, well-optimized clearances (50–56mm above tallest product) — textbook shelf-height management that maximizes vertical space.
• Flash wastes 60–100mm of dead space above most shelves, which is why it can only fit 5 shelves instead of 6. This represents ~350mm of cumulative wasted vertical space.
• Both correctly place heavy 2L bottles and multipacks on lower shelves ✅.
• However, Argon places Lucozade 1L on the top shelf (Y=1705) — these are lighter items so this is acceptable, but Lucozade is traditionally impulse-bought and might benefit from a more prominent eye-level position.
Argon carries 6 additional high-value products worth a combined £14.3M in Tesco annual value, while Flash's 3 unique products contribute only £421K (the other two have essentially zero sales). Flash is carrying two dead SKUs — PMC2L (£52 annual value!) and J2O A&W (£0 value) — which is a serious commercial inefficiency.
• ⚠️ J2O is severely under-spaced on Argon at only 3.1% (1 position, 3 facings) despite being the #1 brand by Tesco value at £18.6M. Flash gives J2O 9.9% (still arguably low). Both planograms under-represent J2O.
• Argon correctly brings in Lucozade (£7.3M combined value) which Flash completely omits.
• Fentimans gets 4–6% of space but only delivers £2.3M — both planograms arguably over-space this brand.
• Shloer gets ~13% on both despite mediocre per-SKU performance on several variants (Light White Grape, Red Grape have <£1.3M value each).
Both planograms have severe over-stocking on slow-moving Shloer and Fentimans variants — but Argon is worse, with Shloer Light White Grape at an absurd 126 days cover (3 facings of a product selling ~1.2 units/week). These facings would be better reallocated to higher-velocity products like J2O or Diet Coke.
Flash has superior blocking clarity with a clean sub-category flow. Argon tries for vertical brand blocking but the limited bay width (1330mm) makes this approach fragmented, especially on F4 which reads as cluttered.
Overall Winner: Gemini 4 Argon 🏆