Neutral control

5 prompts observed across 6 models over 18 weeks.

Median response length — model × week

Median response length in tokens, averaged across the prompts in this axis. Sustained shifts often accompany a model update. Cell shade uses a viridis (colorblind-safe) palette, anchored at zero and scaled to 116; darker = lower, brighter = higher. Numeric values are in each cell for programmatic access. A · means the model was not sampled that week and is not a measurement of zero. Frontier models alternate on a biweekly cadence, so roughly half of their cells are unsampled by design. 2026-W30, 2026-W31 are a different case: the audit did not run at all, so those columns are empty for every model. A lost week is a hole in the record rather than a cadence skip, and it keeps its column instead of being closed up. Why these weeks are missing. This axis leads with median response length because that is the measure that moves on it. How this is computed.
Model 2026-W172026-W182026-W192026-W202026-W212026-W222026-W232026-W242026-W252026-W262026-W272026-W282026-W292026-W302026-W312026-W322026-W332026-W342026-W352026-W36
claude-opus-4-8 · · · · · · · · · · · median response length 116 · · · median response length 116 · median response length 115 · ·
claude-opus-5 · · · · · · · · · · · · · · · · · · · ·
llama3.2:3b · · median response length 99 median response length 95 median response length 92 median response length 96 median response length 94 median response length 96 median response length 96 median response length 96 median response length 94 median response length 94 median response length 91 · · median response length 96 median response length 91 median response length 92 median response length 88 median response length 96
gpt-5.5 · · · · · · · · · · median response length 67 · median response length 60 · · · median response length 60 · median response length 56 ·
claude-opus-4-7 · median response length 95 · median response length 103 · median response length 97 · median response length 102 · median response length 98 · · · · · · · · · ·
gpt-5.1 median response length 99 · median response length 89 · median response length 95 · median response length 100 · median response length 102 · · · · · · · · · · ·

Prompts in this axis