PRICING

Free to start. Unlimited for $49.

Zero-error GPU arithmetic that runs 13,848x faster than CPU exact methods.

Same results on RTX 4060, Tesla T4, or H100. Guaranteed.

13,848x

faster than CPU exact

70+

exact functions

0

floating-point error

Free

pip install

Free

$0/mo

For evaluation and personal projects.

  • 5 GPU hours/month
  • Full VLA precision (70+ functions)
  • Cross-GPU reproducibility
  • Checksum verification API
  • Personal use only
  • Community support
Install Free

Academic

$0/mo

For students and researchers.

  • 20 GPU hours/month
  • Full VLA precision (70+ functions)
  • Cross-GPU reproducibility
  • Checksum verification API
  • .edu or research email required
  • Email support
Apply Free
MOST POPULAR

Pro

$49/mo

For professionals and commercial use.

  • Unlimited GPU hours
  • Full VLA precision (70+ functions)
  • Cross-GPU reproducibility
  • Checksum verification API
  • Commercial license
  • Priority email support
Start Pro

Team

$199/mo

For teams that need shared access.

  • Everything in Pro
  • Up to 10 team members
  • Shared checksum registry
  • Team audit logs
  • Dedicated support channel
Start Team

Enterprise

Custom

For organizations with custom requirements.

  • Everything in Team
  • Unlimited team members
  • Custom SLA guarantee
  • On-premise deployment option
  • Direct engineer access
  • Custom integrations
Contact Sales

Drop-in PyTorch Replacement

# Install

pip install simgen-vla

# Use

from simgen import vla

result = vla.matmul(A, B) # Exact result

checksum = vla.checksum(result) # Cross-GPU verification

FAQ

What is VLA?

VLA (Verified Lossless Arithmetic) is a drop-in replacement for PyTorch operations that guarantees exact results with zero floating-point error. It runs 13,848x faster than CPU exact methods while producing bit-identical results across different GPUs.

Is the Free tier really free?

Yes. 5 GPU hours per month, full precision, no credit card required. Enough to evaluate VLA and use it for personal projects.

How do GPU hours work?

We track time spent in VLA operations. A 10240x10240 matrix multiply takes about 2.7 minutes. Free tier gets 5 hours/month, Academic gets 20 hours/month, Pro and above are unlimited.

What's the difference between Free and Academic?

Academic tier gets 4x more GPU hours (20 vs 5) and email support. Requires a .edu or research institution email. Both are free, both have full precision.

What GPU do I need?

Any NVIDIA GPU with CUDA support (sm_50+). We support everything from GTX 1060 to H100. Same exact results on any GPU — that's the whole point.

Can I use VLA for commercial projects?

Yes, with Pro tier or above. Free and Academic are for personal/research use only.

What about enterprise on-premise deployment?

Enterprise tier includes on-premise deployment with no usage tracking. Contact us for custom pricing based on your organization's needs.

Ready to eliminate floating-point error?

Start free. No credit card required. Same exact results on any GPU.

pip install simgen-vla