From 16b2b8a23cc9199f734f52193da6c667c95ba83d Mon Sep 17 00:00:00 2001 From: Chenyi An Date: Fri, 10 Jul 2026 17:40:53 +0800 Subject: [PATCH] chore: add foundry skill benchmark for gpt5.6 luna and terra --- doc/microsoft-foundry-skill-status.md | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) diff --git a/doc/microsoft-foundry-skill-status.md b/doc/microsoft-foundry-skill-status.md index d823cad..f3eb418 100644 --- a/doc/microsoft-foundry-skill-status.md +++ b/doc/microsoft-foundry-skill-status.md @@ -34,7 +34,7 @@ The response should mention Foundry agent workflows such as scaffolding, local t ## Current Benchmark -The Foundry skill performs consistently across different Copilot models. The benchmark tested Claude Opus 4.8, Claude Opus 4.6, Claude Sonnet 4.6, GPT-5.5, GPT-5.4, GPT-5.3-Codex, GPT-5 mini, and MAI-Code-1-Flash. +The Foundry skill performs consistently across different Copilot models. The benchmark tested Claude Opus 4.8, Claude Opus 4.6, Claude Sonnet 4.6, GPT-5.6 Luna, GPT-5.6 Terra, GPT-5.5, GPT-5.4, GPT-5.3-Codex, GPT-5 mini, and MAI-Code-1-Flash. Benchmark prompt: @@ -47,6 +47,8 @@ Representative results: | Copilot model | Time Cost | AI Credits | | --- | ---: | ---: | | Sonnet 4.6 | 10 min 30 s | 100 | +| GPT-5.6 Terra | 9 min | 90 | +| GPT-5.6 Luna | 8 min | 45 | | Auto (GPT-5.3-Codex) | 9 min 30 s | 60 | | Free plan / cheapest (GPT-5 mini) | 11 min 30 s | 20 |