Infer Lab Logo
ProductContactLogin
Infer Lab Logo

Pricing

Fair pricing based on actual optimization results. Only pay for the performance improvements you get.

✨ Try first, see results, then pay based on your gains

Pricing Formula

Calculation Steps:

1

Performance Score

(Latency Improvement % + Throughput Improvement %) ÷ 2

Score = (40% + 70%) ÷ 2 = 55.0%
2

Minimum Performance Threshold

Performance Score must be ≥ 10% to qualify for payment

55.0% 10% → Paid ✓
3

Final Price

Model Size (GB) × Performance Score (%) × $0.08

Price = 7 GB × 55.0% × $0.08 = $30.80

💡 Fair Pricing Guarantee:

If optimization improves performance by less than 10%, you pay nothing. We only charge when we deliver real, measurable improvements.

Try the Calculator:

Estimated Price: $30.80

per hardware platform

Simple Pricing Options

Single Platform

Optimize for one hardware platform (Intel, NVIDIA, AMD, or Qualcomm)

Calculated Price

Based on your optimization results

  • PyTorch & ONNX support
  • Full optimization for 1 platform
  • Detailed benchmark report
  • Optimized model download
  • Lifetime access to re-optimize
BEST VALUE

All Platforms Bundle

Optimize for all hardware platforms at once

3.5× Price

Save 12.5% vs 4 individual optimizations

  • Everything in Single Platform
  • All 4 hardware platforms
  • Intel CPU/GPU/NPU optimization
  • NVIDIA GPU optimization
  • AMD CPU/GPU optimization
  • Qualcomm NPU/GPU optimization
  • Cross-platform comparison

Real Pricing Examples

ModelSizeLatency ↓Throughput ↑Performance ScoreSingle PlatformAll Platforms
LLaMA 2 7B7 GB42%68%55.0%$30.80$107.80
LLaMA 2 13B13 GB38%85%61.5%$63.96$223.86
GPT-J 6B6 GB35%55%45.0%$21.60$75.60
LLaMA 2 70B70 GB30%45%37.5%$210.00$735.00

* Actual prices calculated after optimization based on your specific results

How We Measure Performance

LatencyReduction

Improvement = (Baseline - Optimized) / Baseline × 100%

Example:

Baseline: 245ms

Optimized: 89ms

Improvement: (245-89)/245 × 100 = 63.7%

ThroughputIncrease

Improvement = (Optimized - Baseline) / Baseline × 100%

Example:

Baseline: 4.1 tok/s

Optimized: 11.2 tok/s

Improvement: (11.2-4.1)/4.1 × 100 = 173.2%

Ready to Optimize Your Models?

Start optimizing now with full access to benchmarks and results. Pay only based on your actual performance gains.

No credit card required