gpt · 2 arms · 5 effortsgpt-5.4

One model across every effort and arm. Each cell is a run — open it for the full card. Failure chips mark runs with no valid slot.

effort strip

arm \ effort
none
low
medium
high
xhigh
api
gpt-5.4 api noneopen run ↗
39.8K tokens8 runs · all valid
gpt-5.4 api lowopen run ↗
42.9K tokens8 runs · all valid
gpt-5.4 api mediumopen run ↗
102.2K tokens8 runs · all valid
gpt-5.4 api highopen run ↗
368.6K tokens8 runs · all valid
gpt-5.4 api xhighopen run ↗
1M tokens8 runs · all valid
codex-oauth
—
gpt-5.4 codex-oauth lowopen run ↗
44.9K tokens4 runs · all valid
gpt-5.4 codex-oauth mediumopen run ↗
44.7K tokens4 runs · all valid
gpt-5.4 codex-oauth highopen run ↗
64.8K tokens4 runs · all valid
gpt-5.4 codex-oauth xhighopen run ↗
174.8K tokens4 runs · all valid

effort → output size

Output completion tokens per effort, per arm. A read on how much each model spends as effort rises — not a quality axis.

25.4×none → xhigh output
01M2Mnonelowmediumhighxhigh1M39.8K
APIcodex (OAuth)
gpt-5.4 — model · weather-card-benchmark