CogVLM2 19B needs ~28.0 GB VRAM. NVIDIA DGX Spark 128GB has 108.8 GB. With Q4_K_M quantization, expect ~15 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
Fit status
Runs well
Decode
15.2 tok/s
TTFT
12743 ms
Safe context
8K
Memory
28.0 GB / 108.8 GB
This setup is broadly balanced for this model.
Shared-memory contention still exists
The OS, browser, and inference runtime all compete for the same physical memory pool, so real-world headroom is less forgiving than raw capacity suggests.
| Workload | Grade | Fit | Decode | TTFT | Context |
|---|---|---|---|---|---|
| Chat | A | Runs well | 15.2 tok/s | 6951 ms | 8K |
| Coding | A | Runs well | 15.2 tok/s | 12743 ms | 8K |
| Agentic Coding | A | Runs well | 15.2 tok/s | 18535 ms | 8K |
| Reasoning | A | Runs well | 15.2 tok/s | 15060 ms | 8K |
| RAG | A | Runs well | 15.2 tok/s | 23169 ms | 8K |
How CogVLM2 19B (19B params) fits at each quantization level on NVIDIA DGX Spark 128GB (92.2 GB usable).
| Quant | Bits | VRAM | Quality | Fit |
|---|---|---|---|---|
Q2_K | 2 | 7.4 GB | Low | A73 |
Q3_K_S | 3 | 9.3 GB | Low | A73 |
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 99Your hardware
| Model | Params | Grade | Decode | Capabilities |
|---|---|---|---|---|
 Devstral 2 123B Instruct | 123B | S | 2.4 tok/s | |
 Qwen3-Coder 30B A3B Instruct | 30.5B | S |
Yes, NVIDIA DGX Spark 128GB can run CogVLM2 19B with a A grade (Runs well). Expected decode speed: 15.2 tok/s.
CogVLM2 19B (19B parameters) requires approximately 28.0 GB of memory with Q4_K_M quantization.
The recommended quantization for CogVLM2 19B is Q4_K_M, which balances quality and memory efficiency.
On NVIDIA DGX Spark 128GB, CogVLM2 19B achieves approximately 15.2 tokens per second decode speed with a time-to-first-token of 12743ms using Q4_K_M quantization.
For coding workloads, CogVLM2 19B on NVIDIA DGX Spark 128GB receives a A grade with 15.2 tok/s and 8K context.
On NVIDIA DGX Spark 128GB, CogVLM2 19B can safely use up to 8K tokens of context. 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-dgx-spark-128gb" width="400" height="180" frameborder="0" style="border:none;border-radius:12px;overflow:hidden;" title="Will It Run AI — fit result"></iframe>
Preview:
10.6 GB |
| Medium |
| A73 |
Q4_K_M | 4 | 11.6 GB | Medium | A73 |
Q5_K_M | 5 | 13.7 GB | High | A73 |
Q6_K | 6 | 15.6 GB | High | A73 |
Q8_0 | 8 | 20.3 GB | Very High | A74 |
F16Best for your GPU | 16 | 38.9 GB | Maximum | A78 |
| 24.8 tok/s |
| Qwen 3.5 27B | 27B | A | 10.7 tok/s |
| Qwen 3.6 27B | 27B | A | 8.2 tok/s |
| Qwen 3.5 122B A10B | 122B | S | 6.6 tok/s |
Not always. NVIDIA DGX Spark 128GB can often fit larger models thanks to unified memory, but a discrete GPU with dedicated high-bandwidth VRAM may still decode faster once the model fits. For this combination, the important distinction is capacity versus sustained throughput.