SemiAnalysis analyzed 616,300 OpenAI responses: GPT 5.6 Terra averages 1.94 tool calls per reply
AccBalanced · x · 2026-09-18
Following its Claude breakdown, SemiAnalysis analyzed 616,300 OpenAI responses from internal usage to measure tool-call frequency:
- GPT 5.6 Terra: 1.94 client tool calls per response, the highest by far
- Sol: 1.00
- Luna: 0.97
- GPT 6 Astra (early usage, 11,717 responses): 1.18, between Terra and the others
The data reveals large behavioral differences in how actively each OpenAI model leans on tool calls.
More from Models
- ChatGPT co-inventor launches Jev model claiming 20-200x speed and 40-400x cost cuts — multiply_matrix · 2026-09-18
- Matt Holden: structured output is LLMs' real magic, now at 300ms and nearly free — holdenmatt · 2026-09-18
- An AI forecaster has won the seasonal Metaculus Cup for the first time — NathanpmYoung · 2026-09-18
- After Jev's classification model hit, will generic regression and time-series foundation models be next as a service? — hichaelmart · 2026-09-18
- Jev vs generic model at same cost: 82.9% vs 58.8% on MMLU-Pro — simonguozirui · 2026-09-18
- DeepSeek V4.1-Flash vs V4-Pro Benchmarked: 3x Cheaper but Slower and Weaker Output — Arindam_1729 · 2026-09-18