Qwen3 on NVIDIA GeForce RTX 4070 · 12GB VRAM

✅ Runs (tight) — Qwen3 14B @ Q4_K_M
Qwen3 14B at Q4_K_M needs ~10.1 GB (weights 8.3 GB + KV 1.3 GB + overhead 512 MB @ 8K ctx) of your 10.8 GB usable VRAM — fits, but little headroom — close other apps or trim context. Expect ~20–35 tok/s (fast).
context @ Q4_K_M: runs to 64K · won't fit past that on this rig
quantneedsspeed
~ Q8_016.6 GB—
~ Q5_K_M11.5 GB~6–9 tok/s (slow)
✓ Q4_K_M10.1 GB~20–35 tok/s (fast)
✓ Q3_K_M8.5 GB~25–45 tok/s (fast)
$runlocal install qwen3:14b
⚠ NVIDIA support is best-effort in v0.1 — verify before relying on it.
runlocal verdict card
download card share on x share on reddit drop it in your model-drop post
Qwen3: Runs (tight)
for your own README — links back here

Same model, other machines: Apple M1 · Apple M1 Pro · Apple M2 · Apple M3 Pro · Apple M2 Pro · Apple M4 Pro

Also runs on a NVIDIA GeForce RTX 4070: DeepSeek-R1 (distill) · Qwen2.5 · Llama 3.1 · Gemma 4 · Mistral 7B

Best models for a NVIDIA GeForce RTX 4070 · 12GB VRAM →

Architecturally similar

These share Qwen3's KV-cache geometry — they size memory the same way, so here's how they fit on a NVIDIA GeForce RTX 4070 · 12GB VRAM too:

modelsharesverdict here
Qwen2.5-Coder 4kv/128hd8kv/128hd ✅ Runs (tight)
Llama 3.2 8kv/128hd ✅ Runs comfortably
Llama 3.3 8kv/128hd ❌ Won't fit
Mistral Small 3 8kv/128hd ⚠️ Partial GPU offload
Mistral Nemo 8kv/128hd ✅ Runs (tight)
Codestral 8kv/128hd ⚠️ Partial GPU offload
Llama 3.2 Vision 8kv/128hd ✅ Runs comfortably
Mixtral 8x7B 8kv/128hd 🐢 CPU-only (slow)

…and 11 more — see the Qwen3 page.

Check any combo yourself: open the checker →