Can CogVLM2 19B run on AMD Instinct MI250X 128GB?

YES — Runs Great

A80Great

○Estimated from fit model

CogVLM2 19B needs ~27.7 GB VRAM. AMD Instinct MI250X 128GB has 128.0 GB. With Q4_K_M quantization, expect ~232 tok/s.

Runtime: llama.cppCapacity: RoomyBandwidth: HighStack: StandardBottleneck: Balanced

Operating mode

Choose the run profile you care about

Interactive favors responsiveness, while light API and scale-out lean harder on serving readiness. The fit stays the same, but the recommendation lens changes.

Current mode

Balanced

Balanced for general local use. Keeps the ranking neutral across personal and serving workflows.

Capabilities:

Select quantization to explore

Q4_K_M (Medium quality) — 27.7 GB, 231.5 tok/s, Runs well

27.7 GB required128.0 GB available

22% VRAM used

Fit status

Runs well

Decode

231.5 tok/s

TTFT

836 ms

Safe context

Memory

27.7 GB / 128.0 GB

Memory breakdown

Weights11.6 GB

KV Cache2.4 GB

Runtime0.9 GB

Headroom12.8 GB

See how fast it feels

See how fast it feelsCogVLM2 19B on AMD Instinct MI250X 128GB
1st promptCold start — includes initialization
>What is local AI inference?
Local AI inference means running an AI language model directly on your own hardware — your laptop, desktop, or server — instead of sending requests to a remote cloud API.

When you run inference locally the model weights are loaded into your GPU or unified memory. Each token you generate requires reading those weights from memory, so memory bandwidth is the main bottleneck for decode speed.

Key benefits of running locally:
 - Full privacy: your prompts never leave your machine
 - No per-token cost or rate limits
 - Works offline once the model is downloaded
 - Latency depends only on your hardware
2nd promptWarm — model ready, no init overhead
>How much VRAM do I need?
It depends on the model size and quantization level. A rough rule of thumb:

 Model size Q4 (4-bit) Q8 (8-bit) FP16
 7B params ~4.3 GB ~7.5 GB ~14 GB
 13B params ~7.9 GB ~13.9 GB ~26 GB
 70B params ~42.7 GB ~74.9 GB ~140 GB

Most people use 4-bit quantization (Q4_K_M) which gives 90-95% of full quality at a fraction of the memory. A 24 GB GPU can comfortably run most 7B-13B models.
Estimated: 231.5 tok/s decode · 836ms TTFT (warm) · 579 tok/s prefill

What limits this setup

This setup is broadly balanced for this model.

No major red flags

This recommendation has enough memory headroom and acceptable estimated speed for the selected workload.

Best improvement path

Performance by workload

Workload	Grade	Fit	Decode	TTFT	Context
Chat	A	Runs well	231.5 tok/s	456 ms	8K
Coding	A	Runs well	231.5 tok/s	836 ms	8K
Agentic Coding	A	Runs well	231.5 tok/s	1216 ms	8K
Reasoning	A	Runs well	231.5 tok/s	988 ms	8K
RAG	A	Runs well	215.4 tok/s	1634 ms	8K

Quantization options

How CogVLM2 19B (19B params) fits at each quantization level on AMD Instinct MI250X 128GB (128.0 GB usable).

Quant	Bits	VRAM	Quality	Fit
Q2_K	2	7.4 GB	Low	A71
Q3_K_S	3	9.3 GB	Low	A72
NVFP4	4

Get started

Copy-paste commands to run CogVLM2 19B on your machine.

Run

docker run --rm -it ghcr.io/ggerganov/llama.cpp:full \
 --hf-repo "THUDM/cogvlm2-llama3-chat-19B" \
 --hf-file "cogvlm2-llama3-chat-19B-Q4_K_M.gguf" \
 -c 4096 -ngl 99

Your hardware

More models your AMD Instinct MI250X 128GB can run

Model	Params	Grade	Decode	Capabilities
👁 Mistral Devstral 2 123B Instruct	123B	S	36.2 tok/s
👁 Alibaba Qwen3-Coder 30B A3B Instruct	30.5B	S

Frequently asked questions

See all results for AMD Instinct MI250X 128GB See all hardware for CogVLM2 19B

URL: https://willitrunai.com/can-run/cogvlm2-19b-on-instinct-mi250x-128gb

⇱