How powerful is your computer for AI?

Run a model on your computer from this page. It gets 50 graded, timed tasks, then your computer is ranked.

Run the benchmark Play what models built

Using a coding agent? It can run the benchmark for you: the page for agents has what to paste.

New here? Get a first score with a small model.

What models built

Each page runs in a sandbox. Open one, then swipe or use the arrow keys.

Surprise me · Every page

What are you running on?

Hardware type
How do I find this?
  • Mac: Apple menu, then About This Mac.
  • Windows: open Task Manager (Ctrl+Shift+Esc), then Performance (on Windows 10, click More details first). GPU shows the card's name and dedicated memory; with two GPUs, use the NVIDIA or AMD one. CPU shows the processor's name; Memory shows system RAM.
  • Linux: the processor and system RAM are under Settings, then About. For a graphics card's memory:
    • NVIDIA: run nvidia-smi. The figure after the slash on its memory line is the total, in MiB (divide by 1,024 for GB).
    • AMD: run cat /sys/class/drm/card*/device/mem_info_vram_total. It prints the total in bytes (divide by 1,073,741,824 for GB).
    • Intel: install clinfo, then run clinfo | grep 'Global memory size'. It prints the total in bytes (divide by 1,073,741,824 for GB).

Keeps this computer for 30 days.

Top computers

Score = tokens a second times the share of points passed. One row per person and computer. How scores work

The whole leaderboard

Latest runs

The figure is tasks passed out of the run's total.

  1. 44/50 tasks passed
    recursive/flint-v1.0:free HostedOpenRouter · 146 tok/s · 257k tokens · uwu · 2 days ago
  2. 47/50 tasks passed
    particle-ai/particle-1:free HostedOpenRouter · 199 tok/s · 168k tokens · uwu · 2 days ago
  3. 47/50 tasks passed
    nvidia/nemotron-3.5-super-vl:free HostedOpenRouter · 136 tok/s · 205k tokens · uwu · 2 days ago
  4. 34/50 tasks passed
    nvidia/nemotron-3.5-lightning:free HostedOpenRouter · 158 tok/s · 227k tokens · uwu · 3 days ago
  5. 42/50 tasks passed
    nvidia/nemotron-3-ultra-550b-a55b:free HostedOpenRouter · 44.2 tok/s · 111k tokens · uwu · 3 days ago
  6. 38/50 tasks passed
    nvidia/nemotron-3-nano-omni-30b-a3b-reasoning:free HostedOpenRouter · 68.3 tok/s · 206k tokens · uwu · 3 days ago
  7. 35/50 tasks passed
    liquid/lfm-3-1.7b:free HostedOpenRouter · 217 tok/s · 202k tokens · uwu · 3 days ago
  8. 32/50 tasks passed
    liquid/lfm-2.5-2.6b:free HostedOpenRouter · 356 tok/s · 156k tokens · uwu · 3 days ago
  9. 45/50 tasks passed
    dots-studio/dots-3-note-preview:free HostedOpenRouter · 98.2 tok/s · 321k tokens · uwu · 3 days ago
  10. 44/50 tasks passed
    baidu/ernie-5.5-flash-preview:free HostedOpenRouter · 253 tok/s · 285k tokens · uwu · 3 days ago

All runs