Can RX 7600 8GB run Qwen3-14B?

AMD · 8 GB GDDR6 · 288 GB/s bandwidth · Alibaba · 14.8B · model ctx up to 128K

No — Qwen3-14B at Q4 needs ≈ 10.62 GB even at 4K context, beyond the RX 7600 8GB's 8 GB GDDR6. It would only run with heavy CPU/disk offload. Decode at Q4/8K: ≈ 15.5 tok/s (estimated, batch 1). FP16 needs ≈ 33.56 GB — does not fit on this card. Estimates come from memory-bandwidth math; rows tagged measured override estimates. Relative ranking is more reliable than absolute tok/s. Methodology.

Fit & speed by quant and context

Quant	Context	VRAM needed	Fits?	tok/s (decode)
Q4	4K	10.62 GB	No	20.1 est.
Q4	8K default	12.1 GB	No	15.5 est.
Q4	32K	20.98 GB	No	5.1 est.
FP16	4K	32.08 GB	No	—
FP16	8K default	33.56 GB	No	—
FP16	32K	42.44 GB	No	—

Fits = weights + KV(ctx) + 1 GB overhead ≤ 95% of VRAM. Measured anchors are context-agnostic; the fit verdict is recomputed per context. A missing tok/s means the model is far beyond this card (offload-only territory).

VRAM breakdown at 8K context

Quant	Weights	KV cache	Overhead	Total needed	RX 7600 8GB VRAM
Q4	8.14 GB	2.96 GB	1 GB	12.1 GB	8 GB
FP16	29.6 GB	2.96 GB	1 GB	33.56 GB	8 GB

Other GPUs that run Qwen3-14B

• B200 192GB SXM — ≈344.0 tok/s Q4 est.
• MI300X 192GB — ≈227.9 tok/s Q4 est.
• H200 141GB SXM5 — ≈206.4 tok/s Q4 est.
• H100 80GB SXM5 — ≈144.0 tok/s Q4 est.
• Gaudi 3 128GB — ≈140.9 tok/s Q4 est.
• A100 80GB SXM4 — ≈87.7 tok/s Q4 est.
• RTX 5090 32GB — ≈77.1 tok/s Q4 est.
• RTX 4090 24GB — ≈69.1 tok/s Q4 measured

All GPUs for Qwen3-14B →

Other models for the RX 7600 8GB

• Qwen3-0.6B — ≈229.1 tok/s Q4 est.
• gemma-3-1b-it — ≈183.3 tok/s Q4 est.
• Llama-3.2-1B-Instruct — ≈152.7 tok/s Q4 est.
• Qwen3-1.7B — ≈91.6 tok/s Q4 est.
• Llama-3.2-3B-Instruct — ≈57.3 tok/s Q4 est.
• Phi-4-mini-instruct — ≈48.2 tok/s Q4 est.
• Phi-4-mini-reasoning — ≈48.2 tok/s Q4 est.
• gemma-3-4b-it — ≈47.0 tok/s Q4 est.

All models on the RX 7600 8GB →