OpenAI says GPT-6.1 Terra beats Sol on visual understanding; Claude Opus 5.5 tops chart
reach_vb · x · 2026-11-10
Responding to a question, OpenAI's reachvb says GPT-6.1 Terra is more efficient than GPT-6.1 Sol on a variety of visual understanding tasks, citing OpenAI's "Agents' Last Exam v1" score-vs-cost chart:
- GPT-6.1 Terra reaches 56–59% at the lowest cost, best value
- GPT-6.1 Sol costs slightly more with slightly lower scores
- GPT-6 Astra scales to 59% at $8 per run
- Claude Opus 5.5 tops the chart at 63% but at the highest cost
The chart maps score/cost tradeoffs for several GPT variants against Claude Opus 5.5 on agentic visual tasks.
More from Models
- Claude Fable 5.2 scores 64.9% on RoboDojo robot control, dominating rivals — burny_tech · 2026-11-10
- Claude Fable 5.2 video demo wows on pacing, but AI-sounding music persists — jiayuan_jy · 2026-11-10
- Gemini Notebooks keeps pulling context from old conversations, user reports — AnonymousDork929 · 2026-11-10
- User complains model takes four minutes to compare two 16-token completions — cephaloform · 2026-11-10
- DavidAU releases Qwen3.8-40B-Ultimate-Heretic-Uncensored-Turbo-Brilliance-50x-AGI-NEO-GGUF merge — mp3m4k3r · 2026-11-10
- Uncensored Gemma 4 31B 'heretic' build hits Hugging Face, 11.9GB quantized — alexcovo_eth · 2026-11-10