Only with CPU offload
Qwen3 Coder 30B-A3B at Q4_K_M needs 18.5 GB but only 10.6 GB is addressable, so about 46% of the layers would stream from system RAM at roughly 60 GB/s. Expect around 15 tokens per second — usable for batch work, painful for chat.
Agentic coding MoE with a 256K native window. Still the most-downloaded local code model.
Hardware requirements What hardware do I need to run Qwen3 Coder 30B-A3B? →At Q4_K_M with an 8K context: a 24 GB GPU (RTX 4090) or a 32 GB Mac (M4). Minimum, recommended, high-quality and long-context tiers, with all 114 devices checked at every quantisation.
The VRAM budget
Quantisation ladder
| Quant | Weights | Total @ 8K | Max context | Tok/s | Quality | Fit |
|---|---|---|---|---|---|---|
| Q8_0 | 30.2 GB | 31.5 GB | — | ~5.9 | −0.1% ppl | 20.9 GB over |
| Q6_K | 23.3 GB | 24.6 GB | — | ~8.7 | −0.4% ppl | 14.0 GB over |
| Q5_K_M | 20.1 GB | 21.5 GB | — | ~11 | −0.8% ppl | 10.9 GB over |
| Q4_K_M | 17.1 GB | 18.5 GB | — | ~15 | −1.9% ppl | 7.9 GB over |
| Q3_K_M | 13.9 GB | 15.2 GB | — | ~24 | −5.4% ppl | 4.6 GB over |
| Q2_K | 11.9 GB | 13.2 GB | — | ~37 | −15% ppl | 2.6 GB over |
Quality is the published perplexity delta against f16 weights. Max context assumes an f16 KV cache; q8_0 roughly doubles it.
How to run it
$ llama-server \
-hf Qwen/Qwen3-Coder-30B-A3B-Instruct:Q4_K_M \
-c 8192 -ngl 25
The engine underneath most of the others. Every knob is exposed. More on llama.cpp.
Related models on a GeForce RTX 5070
| Model | Total | Tok/s est. | |
|---|---|---|---|
| Qwen3 30B-A3B 30.5B MoE · same family |
Q4_K_M · 18.5 GB | ~15 | CPU offload |
| Qwen3 32B 32.8B · same family |
Q4_K_M · 21.0 GB | ~3.3 | CPU offload |
| Qwen3 14B 14.8B · same family |
Q4_K_M · 10.2 GB | 49 | Tight fit |
| Qwen3 8B 8.19B · same family |
Q4_K_M · 6.3 GB | 88 | Runs great |
| GLM-4.7-Flash 30B-A3B 31.2B MoE · similar size |
Q4_K_M · 18.6 GB | ~16 | CPU offload |
| Muse Glimmer 30B 29.8B · similar size |
Q4_K_M · 17.5 GB | ~4.7 | CPU offload |
| Gemma 4 31B 31.3B · similar size |
Q4_K_M · 20.2 GB | ~3.5 | CPU offload |
Qwen3 Coder 30B-A3B on other hardware
| Device | Total | Tok/s est. | |
|---|---|---|---|
| GeForce RTX 4060 Laptop 7.0 GB usable · 256 GB/s |
18.5 GB | ~10 | CPU offload |
| GeForce RTX 3060 12 GB 10.6 GB usable · 360 GB/s |
18.5 GB | ~14 | CPU offload |
| M4 · 16 GB 10.7 GB usable · 120 GB/s |
18.5 GB | — | Won't fit |
| GeForce RTX 4090 22.4 GB usable · 1008 GB/s |
18.5 GB | 126 | Runs great |
| Radeon RX 7900 XTX 22.4 GB usable · 960 GB/s |
18.5 GB | 120 | Runs great |
| M4 · 32 GB 24.0 GB usable · 120 GB/s |
18.5 GB | 17 | Runs great |
| M3 Max · 64 GB 48.0 GB usable · 400 GB/s |
18.5 GB | 56 | Runs great |