User finds GPT-6.1-Terra unreliable for code reviews, mulls switch to Sol or Astra
alexcovo_eth · x · 2026-11-06
A developer tried GPT-6.1-Terra for implementation reviews and reports it performs poorly, saying they can't rely on it at all. They're considering switching back to Sol or using Astra for implementation work, and wonder if Terra is better suited for creative tasks.
More from Models
- API dies mid-task: only Atomic Agent with local fallback aced the test, Codex just hung — Ok-Brush8377 · 2026-11-06
- 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