Skip to content
ModelPerf
.ai
Leaderboard
Analysis
ROI calculator
Methodology
Custom benchmarks
Pricing
Sign in
Estimator
Benchmark estimator
Performance, energy and operating costs for your model and hardware.
Modeled estimate
Model
Step 1 of 3
Select model
Upload weights
Model
DeepSeek-V4-Pro
DeepSeek-V4.1-Flash
gemma-4-31B-it
gemma-4-E4B-it
GLM-5.3
GLM-5.3-Flash
Mistral-Medium-3.5-128B
Mistral-Small-4-119B-2603
Qwen3.8-27B
Parameters (B)
Hardware
Step 2 of 3
NVIDIA B200 x8
$48/hr
·
8.0 kW
NVIDIA H200 x8
$32/hr
·
5.6 kW
NVIDIA H100 x8
$22/hr
·
5.6 kW
AMD MI355X x8
$30/hr
·
9.6 kW
AMD MI300X x8
$20/hr
·
6.0 kW
TPU v6e x8
$22/hr
·
3.2 kW
AWS Trainium2
$25/hr
·
4.0 kW
Operating costs
Step 3 of 3
Electricity ($/kWh)
Datacenter PUE
Cooling ($/month)
Network ($/month)
Staff ($/month)
Other ($/month)
Utilization (%)
Generate estimate