CogVLM2 19B needs ~11.8 GB VRAM. RTX 3080 10GB has 10.0 GB. With Q2_K quantization, expect ~34 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
5.9 GB over capacity — needs offload or smaller quantization
Fit status
Too heavy
Decode
13.5 tok/s
TTFT
14331 ms
Safe context
4K
Memory
15.9 GB / 10.0 GB
Offload
40%
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 1.1 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 | 16.0 tok/s | 6609 ms | 4K |
| Coding | F | Too heavy | 13.5 tok/s | 14331 ms | 4K |
| Agentic Coding | F | Too heavy | 10.0 tok/s | 28142 ms | 4K |
| Reasoning | F | Too heavy | 13.5 tok/s | 16937 ms | 4K |
| RAG | F | Too heavy | 10.0 tok/s | 35177 ms | 4K |
How CogVLM2 19B (19B params) fits at each quantization level on RTX 3080 10GB (10.0 GB usable).
| Quant | Bits | VRAM | Quality | Fit |
|---|---|---|---|---|
Q2_K | 2 | 7.4 GB | Low | F0 |
Q3_K_S | 3 | 9.3 GB | Low | F0 |
NVFP4 | 4 |
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 options
RTX 5060 Ti 16GBBudget pickMakes 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.
~$449 MSRP
RTX 4060 Ti 16GBBest valueMakes 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.
~$499 MSRP
RTX 2000 Ada 16GBNVIDIA upgradeMakes 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.
~$625 MSRP
RTX 5090 Laptop 24GBBiggest leapMakes 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, RTX 3080 10GB can run CogVLM2 19B at Q2_K quantization (Very compromised (needs ~1.1 GB host RAM)). The recommended Q4_K_M requires 15.9 GB which exceeds available memory, but at Q2_K it needs only 11.8 GB. Expected decode speed: 34.1 tok/s.
CogVLM2 19B (19B parameters) requires approximately 15.9 GB at Q4_K_M quantization. On RTX 3080 10GB, it fits at Q2_K using 11.8 GB.
The recommended quantization is Q4_K_M, but on RTX 3080 10GB the best fitting quantization is Q2_K, which uses 11.8 GB.
On RTX 3080 10GB, CogVLM2 19B achieves approximately 34.1 tokens per second decode speed with a time-to-first-token of 5681ms using Q2_K quantization.
For coding workloads, CogVLM2 19B on RTX 3080 10GB receives a F grade with 13.5 tok/s and 4K context.
On RTX 3080 10GB, CogVLM2 19B can safely use up to 5K tokens of context at Q2_K quantization. The model's official context limit is 8K, but available memory constrains the safe maximum.
Paste this snippet into any page to show a live fit card.
<iframe src="https://willitrunai.com/embed/cogvlm2-19b-on-rtx-3080-10gb" width="400" height="180" frameborder="0" style="border:none;border-radius:12px;overflow:hidden;" title="Will It Run AI — fit result"></iframe>
Preview:
| 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 |
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.