Can RTX 5070 12GB run gpt-oss-20b?

NVIDIA · 12 GB GDDR7 · 672 GB/s bandwidth · OpenAI · 21.5B (active 3.6B) · MoE · model ctx up to 128K

No — gpt-oss-20b at Q4 needs ≈ 13.19 GB even at 4K context, beyond the RTX 5070 12GB's 12 GB GDDR7. It would only run with heavy CPU/disk offload. Decode at Q4/8K: ≈ 266.2 tok/s (estimated, batch 1). FP16 needs ≈ 44.72 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	13.19 GB	No	280.9 est.
Q4	8K default	13.55 GB	No	266.2 est.
Q4	32K	15.71 GB	No	198.0 est.
FP16	4K	44.36 GB	No	6.8 est.
FP16	8K default	44.72 GB	No	6.7 est.
FP16	32K	46.88 GB	No	6.1 est.

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	RTX 5070 12GB VRAM
Q4	11.83 GB	0.72 GB	1 GB	13.55 GB	12 GB
FP16	43.0 GB	0.72 GB	1 GB	44.72 GB	12 GB

Other GPUs that run gpt-oss-20b

• B200 192GB SXM — ≈1414.1 tok/s Q4 est.
• MI300X 192GB — ≈936.9 tok/s Q4 est.
• H200 141GB SXM5 — ≈848.5 tok/s Q4 est.
• H100 80GB SXM5 — ≈592.2 tok/s Q4 est.
• Gaudi 3 128GB — ≈579.1 tok/s Q4 est.
• A100 80GB SXM4 — ≈360.4 tok/s Q4 est.
• RTX 5090 32GB — ≈282.5 tok/s Q4 measured
• RTX 4090 24GB — ≈225.2 tok/s Q4 measured

All GPUs for gpt-oss-20b →

Other models for the RTX 5070 12GB

• Qwen3-0.6B — ≈534.5 tok/s Q4 est.
• gemma-3-1b-it — ≈427.6 tok/s Q4 est.
• Llama-3.2-1B-Instruct — ≈356.4 tok/s Q4 est.
• Qwen3-1.7B — ≈213.8 tok/s Q4 est.
• Llama-3.2-3B-Instruct — ≈133.6 tok/s Q4 est.
• Mistral-7B-Instruct-v0.3 — ≈120.5 tok/s Q4 measured
• Phi-4-mini-instruct — ≈112.5 tok/s Q4 est.
• Phi-4-mini-reasoning — ≈112.5 tok/s Q4 est.

All models on the RTX 5070 12GB →