GPT-5.6 Terra
A mid-tier model in the GPT-5.6 series designed for high-frequency everyday workloads. Its performance matches GPT-5.5 at a lower price point, and its Coding Agent Index score is slightly higher than that of Fable 5.
Timeline (last 90 days)
Swapping Terra for Luna is a complete change of magnitude: GitHub's model and pricing table classifies Sol in the top category, Terra in Versatile, and Luna in Lightweight, and Luna's output price is only one twenty-fourth of Terra's ($0.50 vs $12).
Starting November 16, 2026, three models—Claude Opus 4.8, Claude Opus 5, and GPT-5.6 Terra—will be removed from GitHub Copilot's Chat, inline editing, ask and agent modes, and code completion.
GLM-5.3-Flash 在 AutomationBench 上得48.8分，领先 Claude Opus 4.8 的41.0分和 GPT-5.6 Terra 的37.2分，单任务成本约0.045美元。
GLM-5.3-Flash 在 Terminal-Bench 2.1 得84.3分，略低于 Claude Opus 4.8 的85.0分和 GPT-5.6 Terra 的87.4分。
46个模型中有41个完整跑完全部28道题，3个因服务错误未完成，GPT-5.6 Terra因配额限制止步24/28，GPT-6 Astra只跑了1道题
Luna's output price dropped from $6 per million tokens to $0.50 within two months, a reduction of more than 90%.
In terms of input token counts, both Astra and Luna used 16 tokens, while Sol and Terra used 26 tokens.
Qwen 3.8 27B surpasses GPT-5.6 Terra, DeepSeek v4 Pro, GPT-5.6 Luna, and Gemini 3.7 Flash on the Agentic index.
According to Artificial Analysis data, Qwen 3.8's 27 billion parameter model scored 51 on the Agentic Index, surpassing GPT-5.6 Terra, DeepSeek v4 Pro, GPT-5.6 Luna, and Gemini 3.7 Flash.
All Facts (7)
Swapping Terra for Luna is a complete change of magnitude: GitHub's model and pricing table classifies Sol in the top category, Terra in Versatile, and Luna in Lightweight, and Luna's output price is only one twenty-fourth of Terra's ($0.50 vs $12).
50%UnverifiedGLM-5.3-Flash 在 AutomationBench 上得48.8分，领先 Claude Opus 4.8 的41.0分和 GPT-5.6 Terra 的37.2分，单任务成本约0.045美元。
50%UnverifiedGLM-5.3-Flash 在 Terminal-Bench 2.1 得84.3分，略低于 Claude Opus 4.8 的85.0分和 GPT-5.6 Terra 的87.4分。
50%Unverified46个模型中有41个完整跑完全部28道题，3个因服务错误未完成，GPT-5.6 Terra因配额限制止步24/28，GPT-6 Astra只跑了1道题
50%UnverifiedIn terms of input token counts, both Astra and Luna used 16 tokens, while Sol and Terra used 26 tokens.
50%UnverifiedQwen 3.8 27B surpasses GPT-5.6 Terra, DeepSeek v4 Pro, GPT-5.6 Luna, and Gemini 3.7 Flash on the Agentic index.
50%UnverifiedStarting November 16, 2026, three models—Claude Opus 4.8, Claude Opus 5, and GPT-5.6 Terra—will be removed from GitHub Copilot's Chat, inline editing, ask and agent modes, and code completion.
50%