Can RX 7600 8GB run Qwen3.5-27B?

AMD · 8 GB GDDR6 · 288 GB/s bandwidth · Alibaba · 27.0B · model ctx up to 256K

No — Qwen3.5-27B at Q4 needs ≈ 18.55 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: ≈ 2.7 tok/s (estimated, batch 1). FP16 needs ≈ 60.4 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	18.55 GB	No	3.6 est.
Q4	8K default	21.25 GB	No	2.7 est.
Q4	32K	37.45 GB	No	—
FP16	4K	57.7 GB	No	—
FP16	8K default	60.4 GB	No	—
FP16	32K	76.6 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	14.85 GB	5.4 GB	1 GB	21.25 GB	8 GB
FP16	54.0 GB	5.4 GB	1 GB	60.4 GB	8 GB

Other GPUs that run Qwen3.5-27B

• B200 192GB SXM — ≈188.6 tok/s Q4 est.
• MI300X 192GB — ≈124.9 tok/s Q4 est.
• H200 141GB SXM5 — ≈113.1 tok/s Q4 est.
• H100 80GB SXM5 — ≈79.0 tok/s Q4 est.
• Gaudi 3 128GB — ≈77.2 tok/s Q4 est.
• A100 80GB SXM4 — ≈48.1 tok/s Q4 est.
• RTX 5090 32GB — ≈42.2 tok/s Q4 est.
• RTX 4090 24GB — ≈23.8 tok/s Q4 est.

All GPUs for Qwen3.5-27B →

Other models for the RX 7600 8GB

• MiniCPM5-1B — ≈169.7 tok/s Q4 est.
• Qwen3.5-2B — ≈91.6 tok/s Q4 est.
• Llama-3.2-3B-Instruct — ≈48.5 tok/s Q4 measured
• Qwen3-4B-Instruct-2507 — ≈45.8 tok/s Q4 est.
• Qwen3-4B-Thinking-2507 — ≈45.8 tok/s Q4 est.
• NVIDIA-Nemotron-3-Nano-4B-BF16 — ≈45.8 tok/s Q4 est.
• gemma-4-E2B-it — ≈35.9 tok/s Q4 est.
• Mistral-7B-Instruct-v0.3 — ≈25.5 tok/s Q4 est.

All models on the RX 7600 8GB →