Gpt 5 6 Intelligence Vs Cost Across Sol Terra Luna
- Source
- artificialanalysis.ai
- Author
- Artificial Analysis articles
- Date
Context
Each OpenAI GPT-5.6 model can run at several reasoning-effort levels, and each level lands at a different point of score and price. Artificial Analysis plotted those points against its Intelligence Index and the average cost of one index task, and found that while each GPT-5.6 model pushes past GPT-5.5 on the Pareto frontier, the edge of options where nothing else is both smarter and cheaper, Luna and Sol are always ahead of Terra (artificialanalysis.ai). In practice, for any Terra effort level there is a Luna or Sol level that is more intelligent at no extra cost, or as intelligent for less. Luna stands out as particularly cost efficient.
The launch gives the endpoints at max effort: Sol scores 59 at $1.04 per task, Terra 55 at $0.55 and Luna 51 at $0.21 (artificialanalysis.ai). The verdict applies to this chart. On the separate Coding Index, which runs models inside an , the launch post has Terra (max) at 77 and Luna (max) at 75.
- eval — A repeatable test for AI quality — a set of tasks plus scoring — used the way software teams use test suites, because model output is too variable to judge by eyeballing.
- AI agent — An AI system that doesn't just answer once but works toward a goal in a loop — taking actions, reading the results, and deciding what to do next.
- agent harness — The scaffolding around a model that turns it into a working agent — the loop, the tools it can call, and the rules for when to stop.
Checking sign-in…
Loading comments…