MathArena
A bird's-eye view of the benchmark: what it measures and every AI IQ chart built on it.
| # | Model | Effort | Score | Expected cost per problem | Completion time (est.) | Weights |
|---|
About MathArena
What MathArena measures
MathArena has source-backed multi-effort configurations. The Cost Efficiency chart and the model table's expandable rows show those configurations; score bars remain one canonical row per model.
Reading the cost view
Each line uses the shared model-configuration dataset to compare source-matched reasoning-effort levels on MathArena. Every point shows the score and expected cost per problem for that exact configuration. Canonical score bars and bell curves remain one point per model.
Reading the completion-time view
Each point is a public model's MathArena score against its estimated completion time (time to last token) for the input and output lengths set by the sliders. Timing comes from the site's per-model speed measurements, not from the benchmark run, so every point is an estimate and labels carry a ~ prefix. The newest generation with a published result is shown by default. Up and to the left is better.
Reading the score bars
Bars rank models by the source-backed benchmark value used for this chart. Longer bars indicate higher published scores.
Reading the model table
Each row is one canonical model: the score shown is the same best published configuration the score bars use. Models with several published reasoning-effort levels expand to one sub-row per level, with the score and expected cost per problem the source reports for that level. Completion time is the site's estimated time to last token for a 10K-token input and 1K-token output (marked ~), not a benchmark run time. The top 20 rows are shown by default; click a column heading to sort, and missing values always sort last.