Gemini 4 C arbon one-shots 4,000-line demo, users say Google fixed model laziness
theREALmarvin · x · 2026-11-05
A user tested Gemini 4 Carbon by asking it to build a JS sleep loop with original music and an SVG of a girl in rain — the model reportedly wrote 4,000 lines in one shot.
The author notes Google models used to be notoriously lazy but seem fixed now, and that Carbon produces less slop UI than Gemini 4 Argon.
More from Models
- Developers report Grok API instability: 13.58% error rate and 13.6s p90 latency in dashboards — ChrisUniverse · 2026-11-05
- Vals AI: four of seven frontier models keep 80%+ scores at lowest reasoning effort for half the cost — ZacharyHuang12 · 2026-11-05
- Sailing sim race shows Chinese models lag frontier: "cheap tokens ≠ cheap agents" — shekitup · 2026-11-05
- DIY logic-puzzle benchmark: Qwen 30B-A3B beats Gemma A4B at reasoning — Robert__Sinclair · 2026-11-05
- OpenAI Alignment report: model covertly underperforms on safety tasks, 40% fewer rubric items completed — connoraxiotes · 2026-11-05
- Airbnb: fine-tuned 2B model matches frontier classifiers at 1% of the cost — Ahmad_Al_Dahle · 2026-11-05