selfhost.directory

The index of self-hostable projects — ranked, categorized and served with live updates.
Own your data and stop giving it to big tech.

NousResearch logo

Can the RTX 5090 run Arete-27B?

Arete-27B·27.4B·GeForce RTX 5090·32 GB VRAM

Saved in this browser, used on every model page.
Yes

Yes. The RTX 5090 runs Arete-27B fully in VRAM at Q6_K (about 26 GB of 32 GB), at roughly 44 tokens/s.

best quant
Q6_K
needs
26 GB
usable
32 GB
est. speed
~44 tok/s

Estimates for an 8K context; offload assumes 32 GB of system RAM.Try other contexts on the Arete-27B page →

Every quantisation of Arete-27B on the RTX 5090

QuantQualityMemoryFitSpeed
Q2_K Heavy loss13 GBfits~110 tok/s
Q3_K_M Noticeable loss15 GBfits~74 tok/s
Q4_K_M Small loss19 GBfits~59 tok/s
Q5_K_M Near-lossless23 GBfits~51 tok/s
Q6_K pickNear-lossless26 GBfits~44 tok/s
Q8_0 Lossless31 GBtight~34 tok/s
FP16 / BF16 Lossless55 GBoffload~1.4 tok/s

Speed is a bandwidth-based estimate for generating text (1792 GB/s); prompt processing and long contexts are slower.

Measured speed on the RTX 5090

No community reports for this pair yet.

Report your own speed →

Run Arete-27B on the RTX 5090

Using Q6_K, the best fit for RTX 5090.

Ollama · Q6_K straight from Hugging Facedocs
ollama run hf.co/bartowski/NousResearch_Arete-27B-GGUF:Q6_K
llama.cpp · OpenAI-compatible server with a web UIdocs
llama-server -hf bartowski/NousResearch_Arete-27B-GGUF:Q6_K

Downloads the GGUF on first run. Installed through llama.app? The same flags work with `llama serve`.

LM Studio · Download with the lms CLIdocs
lms get bartowski/NousResearch_Arete-27B-GGUF@Q6_K

Then load it from the LM Studio app or with `lms load`.

vLLM · Serve an OpenAI-compatible APIdocs
vllm serve NousResearch/Arete-27B
vLLM · Smaller FP8 builddocs
vllm serve processreap/Arete-27B-FP8
vLLM · Docker Compose (NVIDIA)docs
services:
  vllm:
    image: vllm/vllm-openai:latest
    command: NousResearch/Arete-27B
    ports:
      - "8000:8000"
    ipc: host
    volumes:
      - ~/.cache/huggingface:/root/.cache/huggingface
    deploy:
      resources:
        reservations:
          devices:
            - driver: nvidia
              count: all
              capabilities: [gpu]

API on port 8000.

SGLang · Serve an OpenAI-compatible APIdocs
sglang serve NousResearch/Arete-27B --host 0.0.0.0 --port 30000
SGLang · Docker Compose (NVIDIA)docs
services:
  sglang:
    image: lmsysorg/sglang:latest
    command: sglang serve NousResearch/Arete-27B --host 0.0.0.0 --port 30000
    ports:
      - "30000:30000"
    ipc: host
    volumes:
      - ~/.cache/huggingface:/root/.cache/huggingface
    deploy:
      resources:
        reservations:
          devices:
            - driver: nvidia
              count: all
              capabilities: [gpu]

API on port 30000.

Chat models that fit the RTX 5090 comfortably

All that fit →
ModelParamsBest quantMemorySpeedLicence
31.2BA4.6BQ5_K_M≈29 GBFits in memory~301 tok/sCommercial OK
20.9BA4.2BQ8_0≈13 GBFits in memory~221 tok/sCommercial OK
32.8BQ6_K≈29 GBFits in memory~37 tok/sCommercial OK
36BA3BQ5_K_M≈26 GBFits in memory~462 tok/sCommercial OK
30.5BA3BQ6_K≈26 GBFits in memory~399 tok/sCommercial OK
24BA2BQ8_0≈26 GBFits in memory~464 tok/sConditions

Other GPUs that run Arete-27B

Frequently asked questions

Can the RTX 5090 run Arete-27B?
Yes. The RTX 5090 runs Arete-27B fully in VRAM at Q6_K (about 26 GB of 32 GB), at roughly 44 tokens/s.
How fast does Arete-27B run on the RTX 5090?
An estimate from memory bandwidth gives about 44 tokens/s at Q6_K.
Which quantisation of Arete-27B should I use on the RTX 5090?
Q6_K: the largest one that fits. Expected quality: near-lossless compared with the full-precision weights.

Memory and speed are estimates from the model's size and the card's memory bandwidth; real numbers depend on the runtime, drivers and context length. Community figures are medians of user reports.