christiant.io
Christian TaillonThree GPT-5.6 families across every supplied thinking level
Rotate one model space that combines price per benchmark task, Artificial Analysis Intelligence Index, and output tokens per task. The comparison includes GPT-5.6 Sol, Luna, and Terra at None, Low, Medium, High, Xhigh, and Max, plus gpt-oss-120b Low and High.
Drag to rotate, scroll to zoom, and select a point for a direct comparison.
All 20 supplied configurations
Rows are ordered by intelligence. Click a row to select the same point in the visualization. Measured, calculated, and extrapolated per-task values are labeled separately.
| Configuration | Reasoning | Intelligence | Price / task | Output tokens / task | Intel / $ | Speed | Per-task status |
|---|
Methodology
The x-axis is cost per Intelligence Index task, the y-axis is the Artificial Analysis Intelligence Index, and the z-axis is output tokens per task. Direct per-task measurements are supplied for GPT-5.6 Sol Max, Luna Max, Terra Max, and both gpt-oss-120b settings.
For GPT-5.6 None through Xhigh, measured family totals are converted into per-task coordinates using the corresponding family Max total-to-task ratios:
Sol: cost × 0.000436427191014 · tokens × 0.000240619515594Luna: cost × 0.000298996791192 · tokens × 0.000151282505301Terra: cost × 0.000388130127292 · tokens × 0.000202805290573The site preserves each source label. Sol lower levels remain marked Extrapolated; Luna and Terra lower levels are marked Calculated; family Max and gpt-oss points are marked Measured.