Deterministic Stack Engine
Best AI Stack for long-context RAG
Context window and input price for stuffing large corpora.
Objective Weight Distribution
Context:40%In $:25%Elo:20%TTFT:15%
Recommended PickScore 82.8 / 100
GPT-5.6 Luna
OpenAI · Closed flagship
context window p95input price per m p81elo p49ttft ms p98
Output: $0.6/1MSWE-bench: 61.8%
View full model fact sheetFull Stack Leaderboard
Ranked alternatives optimized for long-context RAG based on objective benchmark weighting.