Skip to content

Inkling on 8× H200 NVL: VRAM, speed and fit

Inkling at 8-bit runs on the 8× H200 NVL with 4K of context at about 59 tokens a second.

Text generation · Chat · Ollama / LM Studio

Runs comfortably · on 8× H200 NVL · 8-bit (Q8_0) · Ollama / LM Studio · Chat

GPU memory
950 GB
of 1,128 GB
Context
4K
auto
Speed
59 tok/s
Very fast
First reply
0.3 s
2K prompt

Where the memory goes

Model
942 GB
Context
1.0 GB
Overhead
6.2 GB
OS
0.0 GB
Free
178 GB
Memory
950 GB of 1,128 GB
Context
4K tokens · auto
Speed
59 tok/s · very fast
First reply
0.3 s for a 2K prompt
Running cost
$65/mo at home · $2,325/mo on a cloud 8× MI300X

Validation: Validated

What Inkling on 8× H200 NVL was checked against
CheckTargetResultStatus
Fits in GPU memory1,128 GB usable950 GB neededPass
Headroom5% free178 GB free (16%)Pass
Context for chat4K or more4KPass
Writing speed15 tok/s or more59 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
Inkling
Parameters
952B, 40.9B active
Precision
8-bit (Q8_0)
Context
4K, auto
Chats at once
1
Use
Chat
Runs in
Ollama / LM Studio
Hardware
8× H200 NVL
GPU can use
1,128 GB of 1,128 GB

Running cost

At home
$65/mo
2960 W at the wall
Renting instead
$2,325/mo
8× MI300X at $19/h

4 h a day, power at $0.18 per kWh, cloud at RunPod on-demand prices as of Nov 22, 2026.

How to run it

llama.cpp
llama-server -hf unsloth/inkling-GGUF
vLLM
vllm serve thinkingmachines/Inkling

Inkling on similar machines

All machines: Inkling VRAM requirements. Everything that runs here: what can the 8× H200 NVL run?

Other models that run on 8× H200 NVL

Questions

Can Inkling run on 8× H200 NVL?

Inkling at 8-bit runs on the 8× H200 NVL with 4K of context at about 59 tokens a second.

Report VR-0030F94, calculated Nov 22, 2026 from published specs, not benchmarked.