Qwen3-VL-8B-Instruct on RTX 4070: Runs comfortably (Sep 2026)
Conditional memory judgement for one configuration. Change context, quant, or engine and the conclusion can change. Configuration used on this page: GGUF Q4_K_M (Qwen), 8,192 cached tokens, 1 sequence, engine llama.cpp. Status: Likely fits. Software: Not verified..
Verdict on GeForce RTX 4070 12GB (desktop, GDDR6X)
Yes, run Q6_K. The calculator puts it at 7.5 GiB–8.1 GiB at 8,192 tokens, inside 11.0 GiB of usable VRAM (after the driver reserve). Q8_0 is the next step up and does not clear the bar (9.4 GiB–10.0 GiB).
With Q6_K, the longest context that still counts as a fit is about 20,489 tokens of a 262,144-token native window, one sequence, fp16 KV cache.
| GGUF quant | File weights | KV cache | Estimated total | Verdict |
|---|---|---|---|---|
| Q8_0 | 8.1 GiB | 1.1 GiB | 9.4 GiB–10.0 GiB | Tight |
| Q6_K (recommended) | 6.3 GiB | 1.1 GiB | 7.5 GiB–8.1 GiB | Fits |
| Q5_K_M | 5.4 GiB | 1.1 GiB | 6.7 GiB–7.3 GiB | Comfortable |
| Q4_K_M | 4.7 GiB | 1.1 GiB | 5.9 GiB–6.5 GiB | Comfortable |
Engine llama.cpp, one sequence, fp16 KV. File weights are the published GGUF sizes in the catalog. Totals come from the same calculator that powers the tool below; they are estimates, not a measurement of this setup.
Fixed page configuration
Qwen3-VL-8B-Instruct GGUF Q4_K_M on the catalog RTX 4070 12 GB SKU, the same quant class as the Qwen3-8B allowlisted combo.
- ModelQwen3-VL-8B-Instruct
- HardwareGeForce RTX 4070 12GB (desktop, GDDR6X)
- QuantizationGGUF Q4_K_M (Qwen)
- Context length (cached tokens)8,192
- Engine
llama.cpp - Concurrent sequences1
- KV cache typefp16
Technical identifiers and share data
- Model
qwen3-vl-8b-instruct - Hardware
nvidia-geforce-rtx-4070-12gb-desktop - Quantization
gguf-q4_k_m - Share fragment for this identity
v2.eyJzY2hlbWFWZXJzaW9uIjoyLCJjYWxjdWxhdG9yVmVyc2lvbiI6IjAuMy4wIiwiZGF0YXNldFZlcnNpb24iOiIyMDI2LTEwLTAyLjEiLCJjb25maWciOnsibW9kZWxJZCI6InF3ZW4zLXZsLThiLWluc3RydWN0IiwiaGFyZHdhcmVJZCI6Im52aWRpYS1nZWZvcmNlLXJ0eC00MDcwLTEyZ2ItZGVza3RvcCIsImRldmljZUNvdW50IjoxLCJxdWFudGl6YXRpb25JZCI6ImdndWYtcTRfa19tIiwiY29udGV4dFRva2VucyI6ODE5MiwiY29uY3VycmVudFNlcXVlbmNlcyI6MSwiZW5naW5lIjoibGxhbWEuY3BwIiwia3ZUeXBlIjoiZnAxNiJ9fQ
Memory breakdown
- Weight artifact
- 4.68 GiB (5,027,784,800 bytes)
- Raw KV cache
- 1.13 GiB (1,207,959,552 bytes)
- Engine extras / workspace
- 0.13 GiB–0.67 GiB
- Total estimated demand
- 5.93 GiB–6.48 GiB
- Net calculator budget
- 11.00 GiB
- OS / display reserve
- 1.00 GiB
- Remaining vs net budget
- 4.52 GiB–5.07 GiB
6.48 GiB high estimate vs 11.00 GiB net
Enough memory does not mean this engine, quant, and OS will run. Support does not mean it will be fast.
Other published files on this device
Same model, device and context (8,192 tokens, 1 sequence, FP16 KV). Each row is one published weight-file set with its usual engine, sorted by memory.
| Quantization | Engine | Estimated memory | Fit |
|---|---|---|---|
| BF16 safetensors | vllm | 18.08 GiB–19.99 GiB | Does not fit |
| GGUF Q8_0 (Qwen) | llama.cpp | 9.36 GiB–9.98 GiB | Tight fit |
| GGUF Q6_K (Unsloth) | llama.cpp | 7.51 GiB–8.09 GiB | Fits |
| GGUF Q5_K_M (Unsloth) | llama.cpp | 6.70 GiB–7.26 GiB | Runs comfortably |
| GGUF Q4_K_M (Qwen) This page | llama.cpp | 5.93 GiB–6.48 GiB | Runs comfortably |
Assumptions that affect this result
- Requested context is within the recorded native window of 262144 tokens.
- Full-resident weight estimates use total parameters, including inactive MoE experts.
- Weight bytes come from the selected weight files only, not from every file in the repository.
- Download file size is not peak GPU memory. Runtime layout, allocator padding, and KV are extra.
- Counted files: Qwen3VL-8B-Instruct-Q4_K_M.gguf.
- This checkpoint has a vision component. Text-only inference does not include vision encoder or projector bytes.
- Context budget is the number of cached tokens per sequence (prompt plus reserved generation).
- Concurrent sequences share one model replica; weights are not multiplied by sequence count.
Related pages
Sources
- Qwen3-VL-8B-Instruct model card
- config.json — revision 0c351dd01ed87e9c1b53cbc748cba10e6187ff3b
- safetensors index — metadata.total_size=17534247392
- Artifact repo Qwen/Qwen3-VL-8B-Instruct-GGUF
- Artifact repo unsloth/Qwen3-VL-8B-Instruct-GGUF
- NVIDIA GeForce RTX 40 Series specs — Original 12 GB GDDR6X SKU, not the later GDDR6 variant
Calculator
This combo page keeps the route model and device. Customize to change them on the main calculator; the share then keeps that identity.
Loading catalog snapshot…
Search or pick a catalog checkpoint to see memory demand.