Can RTX 5090 32GB run Nanbeige4.3 33B?
Estimated speed and memory fit for Nanbeige4.3 33B (33.12B dense) on RTX 5090 32GB. We have no measured run for this pair yet, so the speed is an estimate; submit yours.
Short answer
Yes. Nanbeige4.3 33B Q5_K_M fits fully on RTX 5090 32GB at 16k context, using 28.6 of 31.0 GB usable, and generates about ~64 tokens/s (estimated, not measured). For Q5_K_M, our bandwidth-only estimate was ~39 tok/s against a measured median of ~64 (38% too pessimistic).
Measured generation speed by quantization
No measured rows for Nanbeige4.3 33B on RTX 5090 32GB yet. The ~64 tok/s above is a bandwidth estimate (1792 GB/s × 0.71 efficiency). Have this machine? Submit a llama-bench result and this page will switch to measured data.
Generation = tg128 (128 generated tokens, short context). Public rows are medians of community llama.cpp results (sources); our own runs use llama-bench on a rented GPU.
Alternatives
Faster for Nanbeige4.3 33B
- AMD Ryzen AI Max+ 395 (Strix Halo) 128GB: Q4_K_M, 108 GB usable, ~75 tok/s scaled from measured
Other models on RTX 5090 32GB, similar size
- Laguna XS 2.1 (hardware page): Q6_K, fits fully, ~343 tok/s (scaled from measured)
- DeepSeek R1 Distill Qwen 32B (hardware page): Q5_K_M, fits fully, ~246 tok/s (scaled from measured)
- Nemotron 3 Nano 30B-A3B (hardware page): Q5_K_M, fits fully, ~342 tok/s (scaled from measured)
- Nemotron 3.5 Lightning 30B-A3B (hardware page): UD-Q5_K_M, fits fully, ~349 tok/s (measured)
Memory and context limits
Q5_K_M weights are 25.1 GB; the 16k f16 KV cache adds 3.49 GB. Largest context that fits in 31.0 GB usable:
| KV cache type | Max context |
|---|---|
| f16 | 16k |
| q8_0 | 32k |
| q4_0 | 64k |
Native limit 512k. Full KV cache table for Nanbeige4.3 33B.
Copy-paste commands
The advisor's pick for this hardware: Q5_K_M, 16k context.
llama-server -hf bartowski/Nanbeige_Nanbeige4.3-33B-GGUF:Q5_K_M -c 16384 -fa on --jinja -ngl 99
OLLAMA_CONTEXT_LENGTH=16384 ollama serve ollama run hf.co/bartowski/Nanbeige_Nanbeige4.3-33B-GGUF:Q5_K_M
FAQ
How fast is Nanbeige4.3 33B on RTX 5090 32GB?
About 63.8 tokens/s for generation (Q5_K_M, estimated from 1792 GB/s memory bandwidth). Speed drops as the context fills; see the depth table on this hardware's other pages.
How much memory does Nanbeige4.3 33B need on RTX 5090 32GB?
The advisor's pick, Q5_K_M, is 25.1 GB of weights plus 3.49 GB of KV cache at 16k context (28.6 GB), against 31.0 GB usable.
Is the Nanbeige4.3 33B speed on RTX 5090 32GB measured or estimated?
Estimated. No measured llama.cpp run for this pair is on file yet, so the figure comes from memory bandwidth and our calibration against measured hardware.
What can I use instead of RTX 5090 32GB for Nanbeige4.3 33B?
Nothing with less memory in our list fits it fully. Faster: AMD Ryzen AI Max+ 395 (Strix Halo) 128GB (~75 tok/s).
What is the maximum context for Nanbeige4.3 33B on RTX 5090 32GB?
With an f16 KV cache about 16k tokens, 32k with q8_0 and 64k with q4_0, bounded by the model's own limit.
Related
All models on RTX 5090 32GB · Nanbeige4.3 33B on all hardware · VRAM tier · all measured pages · leaderboard
Compare: RTX 4090 24GB vs RTX 5090 32GB · RTX 3090 24GB vs RTX 5090 32GB · 2x RTX 3090 (48GB total) vs RTX 5090 32GB