Skip to content

Gemma 3 27B IT on H100 SXM: VRAM, speed and fit

Gemma 3 27B IT at 8-bit runs on the H100 SXM with 128K of context at about 73 tokens a second.

Text generation · Chat · Ollama / LM Studio

Runs comfortably · on H100 SXM · 8-bit (Q8_0) · Ollama / LM Studio · Chat

GPU memory
38.7 GB
of 80.0 GB
Context
128K
auto
Speed
73 tok/s
54 when full
First reply
0.3 s
2K prompt

Where the memory goes

Model
27.1 GB
Context
10.4 GB
Overhead
1.1 GB
OS
0.0 GB
Free
41.3 GB
Memory
38.7 GB of 80.0 GB
Context
128K tokens · auto
Speed
73 tok/s · 54 when full
First reply
0.3 s for a 2K prompt
Running cost
$11/mo at home · $61/mo on a cloud 2× RTX A4500

Validation: Validated

What Gemma 3 27B IT on H100 SXM was checked against
CheckTargetResultStatus
Fits in GPU memory80.0 GB usable38.7 GB neededPass
Headroom5% free41.3 GB free (52%)Pass
Context for chat16K or more128KPass
Writing speed15 tok/s or more73 tok/sPass
First replyunder 3 s, 2K prompt0.3 sPass

Targets: a reply you can read along with, a first reply the use case tolerates, and 5% of memory left free. Calculated from published specs, not benchmarked.

The setup

Model
Gemma 3 27B IT
Parameters
27.4B
Precision
8-bit (Q8_0)
Context
128K, auto
Chats at once
1
Use
Chat
Runs in
Ollama / LM Studio
Hardware
H100 SXM
GPU can use
80.0 GB of 80.0 GB

Running cost

At home
$11/mo
500 W at the wall
Renting instead
$61/mo
2× RTX A4500 at $0.50/h

4 h a day, power at $0.18 per kWh, cloud at RunPod on-demand prices as of Nov 21, 2026.

How to run it

llama.cpp
llama-server -hf unsloth/gemma-3-27b-it-GGUF
vLLM
vllm serve google/gemma-3-27b-it

Gemma 3 27B IT on similar machines

  • 2× RTX PRO 6000 BlackwellRuns
  • 4× RTX PRO 6000 BlackwellRuns
  • 8× RTX PRO 6000 BlackwellRuns
  • A100 80 GB SXMRuns
  • 4× H200 NVLRuns
  • 8× H200 NVLRuns

All machines: Gemma 3 27B IT VRAM requirements. Everything that runs here: what can the H100 SXM run?

Other models that run on H100 SXM

Questions

Can Gemma 3 27B IT run on H100 SXM?

Gemma 3 27B IT at 8-bit runs on the H100 SXM with 128K of context at about 73 tokens a second.

Report VR-0XPX6JY, calculated Nov 22, 2026 from published specs, not benchmarked.