OpenAI staffer explains GPT-6.1 Terra: midrange bet on vision and technical communication
goodside · x · 2026-10-27
A user claiming to work at OpenAI explained the logic behind GPT-6.1 Terra on Hacker News:
- Model ordering: Model size largely determines performance, per-token cost, and recall of esoteric facts, with the lineup ranked Luna < Terra < Sol < Astra. Pick Astra for pure quality, Sol for moderate price sensitivity, Terra/Luna for tight budgets.
- Size isn't everything: Later-trained models can beat bigger ones on some axes thanks to new data or architectures — e.g. Sonnet 5.5 outperforming Fable 5.1 in some ways, Gemini 3.8 Flash beating 3.1 Pro.
- Where Terra fits: Probably better than Astra and Sol at technical communication, similar on vision and computer use while faster and cheaper, slightly worse elsewhere. A midrange API option — much cheaper than Sol, much smarter than Luna.
- For subscribers: Not a bombshell release; mainly useful if you hit usage limits or want clearer communication and faster computer use.
Related event: OpenAI Launches GPT-6.1 Terra for Computer Use at Half the Price(42 posts)→
More from Models
- Reflection AI's Beam-1.0 tops US open-weight models, hitting 88% of Qwen3.8-2.4T with 23B active params — ryanshrout · 2026-10-27
- GPT-6 Pro reruns crime research: GTA V's 2013 launch tied to violence dips in four regions — emollick · 2026-10-27
- Dev unfreezes MathTutor project: Opus 5.5 finally makes lessons good enough — lucasmeijer · 2026-10-27
- Why temperature 0 isn't deterministic: 1,000 runs of Qwen3-235B yield 80 outputs — JensHonack · 2026-10-27
- EMNLP study: LLMs attribute biased text to groups, refusing only 0.4% of the time — ZhijingJin · 2026-10-27
- OpenAI launches GPT-6.1 Terra, a cheaper model built for fast computer use — sama · 2026-10-27