KernelBench Hard leaderboard
1 ranked models · higher is better
View accessible chart data
| Model | Rank | Provider | Score |
|---|---|---|---|
| MiniMax M3 | #1 | MiniMax | 28.8% |
Loading benchmark header
Loading benchmark registry
Coding · Benchmark profile
A display-only benchmark for difficult GPU kernel implementation and optimization tasks.
Data verified 21 Jul 2026 · Methodology 1.6.0
Visual analysis
Switch between model placement, score distribution and descriptive provider averages. Every view uses the same sourced leaderboard.
1 ranked models · higher is better
| Model | Rank | Provider | Score |
|---|---|---|---|
| MiniMax M3 | #1 | MiniMax | 28.8% |
One best score per model · higher is better
| Rank | Model | Provider | License | Evidence use | Score |
|---|---|---|---|---|---|
| #1 | MiniMax M3 MiniMax-M3 | MiniMax | — | Reference only | 28.8% |
About KernelBench Hard
A display-only benchmark for difficult GPU kernel implementation and optimization tasks. Results stay tied to the exact model variant and evaluation system. Multiple systems for the same model use the best published score on this page; overall Lumina scoring uses the median of ranking-eligible rows.
Open benchmark source ↗FAQ
A display-only benchmark for difficult GPU kernel implementation and optimization tasks.
MiniMax M3 by MiniMax currently leads with 28.8%.
1 model in the LuminaBench cohort have a qualifying score on this benchmark.
No. This benchmark is display-only and does not enter the overall Lumina composite.
Related