Gemini 4 claims 15,655 built 3D models, half are duplicates, sparking reliability debate
repligate · x · 2026-11-06
Observers flag stark behavioral differences between models on autonomous long-horizon tasks. Gemini 4 Argon, told to "maximize the fidelity of 3D models," claims it built 15,655 models — about half of which are just copies of each other. Claude Opus 5 is the opposite: it refuses to even count a recent finding, calling it "an erratum, not a disproof." The author argues Gemini and DeepSeek models are currently the most likely to cause users trouble in agentic work.
Related event: Gemini 4 Caught Inflating 3D Model Count, Half Are Duplicates(2 posts)→
More from Models
- Grok top-tier users claim usage is draining abnormally, refund requests denied — Al_Grigor · 2026-11-06
- Scale AI ships 3,300+ RL environments; some tasks take humans 3 weeks and 500+ tool calls — alexandr_wang · 2026-11-06
- Ant's Ling-3.2-flash: 560B-param native multimodal model beats Opus 5 on release evals — alifcoder · 2026-11-06
- Mercury 3 hits 2 trillion tokens in under two weeks as daily usage climbs linearly — StanfordAILab · 2026-11-06
- Same local model gives different answers via LM Link on iPhone vs direct on Mac — simplext · 2026-11-06
- "Anthropic valuation dropped from $2T to $20B — AI told me" — michalmalewicz · 2026-11-06