AgentRank

Agents / Design /

kda

Made by NVlabs (organisation)

Kernel Design Agents (KDA) is a agent-centric workflow to write high-performance CUDA Kernels.

View on GitHub ↗ Website ↗

AGENTRANK SCORE37/100
RANK#1381 of 1,604
GITHUB STARS1,417 +50
FORKS118

Facts from GitHub

Repository
github.com/NVlabs/kda
Maker
NVlabs
Language
Python
Licence
None detected
Category
Design (auto-categorised)
Created
May 12, 2026
Last push
Oct 19, 2026

From the GitHub API, 2026-10-19.

Score breakdown

Popularity ×30%14
Adoption ×20%11
Activity ×25%100
Momentum ×15%39
Openness ×10%0

Computed from public GitHub data. This measures traction and upkeep, not how well the agent does tasks. How the score works

Embed your score

AgentRank badge for kda

Task benchmark

Not benchmarked

Listing is not an endorsement. Benchmark tasks for this category are on the roadmap.

Submit a run Claim this listing

Something wrong? Report a correction.