CogVLM2 19B needs ~13.9 GB VRAM. RX 6750 XT 12GB has 12.0 GB. With Q3_K_S quantization, expect ~14 tok/s.
Operating mode
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.
Select quantization to explore
4.1 GB over capacity — needs offload or smaller quantization
Fit status
Too heavy
Decode
8.5 tok/s
TTFT
22659 ms
Safe context
4K
Memory
16.1 GB / 12.0 GB
Offload
30%
It fits through host-memory offload, and offload is the main reason performance drops.
CPU or host-memory offload is active
About 10% of the working set spills out of accelerator memory, which usually hurts latency and sustained decode throughput.
Very little memory headroom
You can run the model, but there is not much room left for longer context, bigger batches, extra apps, or future model updates.
Remove offload with more accelerator memory
Prioritize a GPU or unified-memory tier that fits the whole model natively. Removing offload usually helps more than small compute gains.
Buy headroom, not only minimum fit
A slightly larger memory tier gives you safer context growth and makes the recommendation more future-proof.
Increase host RAM if you keep offloading
This setup may need roughly 1.2 GB of extra host RAM just for the offloaded portion, before OS and other tools.
| Workload | Grade | Fit | Decode | TTFT | Context |
|---|---|---|---|---|---|
| Chat | F | Too heavy | 10.1 tok/s | 10473 ms | 4K |
| Coding | F | Too heavy | 8.5 tok/s | 22659 ms | 4K |
| Agentic Coding | F | Too heavy | 6.4 tok/s | 44340 ms | 4K |
| Reasoning | F | Too heavy | 8.5 tok/s | 26778 ms | 4K |
| RAG | F | Too heavy | 6.4 tok/s | 55425 ms | 4K |
How CogVLM2 19B (19B params) fits at each quantization level on RX 6750 XT 12GB (12.0 GB usable).
| Quant | Bits | VRAM | Quality | Fit |
|---|---|---|---|---|
Q2_KBest for your GPU | 2 | 7.4 GB | Low | A85 |
Q3_K_S | 3 | 9.3 GB | Low | F0 |
NVFP4 | 4 | 10.6 GB | Medium | F0 |
Q4_K_M | 4 | 11.6 GB | Medium | F0 |
Q5_K_M | 5 | 13.7 GB | High | F0 |
Q6_K | 6 | 15.6 GB | High | F0 |
Q8_0 | 8 | 20.3 GB | Very High | F0 |
F16 | 16 | 38.9 GB | Maximum | F0 |
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 99Upgrade-Optionen
Makes the model fit on the accelerator instead of staying completely out of reach.
Removes host-memory offload, which is usually the single biggest latency and throughput win.
ca. $329 MSRP
Makes the model fit on the accelerator instead of staying completely out of reach.
Removes host-memory offload, which is usually the single biggest latency and throughput win.
ca. $349 MSRP
Makes the model fit on the accelerator instead of staying completely out of reach.
Removes host-memory offload, which is usually the single biggest latency and throughput win.
ca. $479 MSRP
RTX 5090 Laptop 24GBGrößter SprungMakes the model fit on the accelerator instead of staying completely out of reach.
Removes host-memory offload, which is usually the single biggest latency and throughput win.
Yes, RX 6750 XT 12GB can run CogVLM2 19B at Q3_K_S quantization (Very compromised (needs ~1.2 GB host RAM)). The recommended Q4_K_M requires 16.1 GB which exceeds available memory, but at Q3_K_S it needs only 13.9 GB. Expected decode speed: 13.6 tok/s.
CogVLM2 19B (19B parameters) requires approximately 16.1 GB at Q4_K_M quantization. On RX 6750 XT 12GB, it fits at Q3_K_S using 13.9 GB.
The recommended quantization is Q4_K_M, but on RX 6750 XT 12GB the best fitting quantization is Q3_K_S, which uses 13.9 GB.
On RX 6750 XT 12GB, CogVLM2 19B achieves approximately 13.6 tokens per second decode speed with a time-to-first-token of 14201ms using Q3_K_S quantization.
For coding workloads, CogVLM2 19B on RX 6750 XT 12GB receives a F grade with 8.5 tok/s and 4K context.
On RX 6750 XT 12GB, CogVLM2 19B can safely use up to 4K tokens of context at Q3_K_S quantization. The model's official context limit is 8K, but available memory constrains the safe maximum.
Remove offload with more accelerator memory. Prioritize a GPU or unified-memory tier that fits the whole model natively. Removing offload usually helps more than small compute gains.
Paste this snippet into any page to show a live fit card.
<iframe src="https://willitrunai.com/embed/cogvlm2-19b-on-rx-6750-xt-12gb" width="400" height="180" frameborder="0" style="border:none;border-radius:12px;overflow:hidden;" title="Will It Run AI — fit result"></iframe>
Preview: