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 Tesla T4 run Arete-27B?

Arete-27B·27.4B·Tesla T4·16 GB VRAM

Saved in this browser, used on every model page.
Just about

Just about. At Q3_K_M it needs about 15 GB of the 16 GB available, so keep the context short, at roughly 13 tokens/s.

best quant
Q3_K_M
needs
15 GB
usable
16 GB
est. speed
~13 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 Tesla T4

QuantQualityMemoryFitSpeed
Q2_K Heavy loss13 GBfits~20 tok/s
Q3_K_M pickNoticeable loss15 GBtight~13 tok/s
Q4_K_M Small loss19 GBoffload~6.1 tok/s
Q5_K_M Near-lossless23 GBoffload~4 tok/s
Q6_K Near-lossless26 GBoffload~3 tok/s
Q8_0 Lossless31 GBoffload~2 tok/s
FP16 / BF16 Lossless55 GBtoo big—

Speed is a bandwidth-based estimate for generating text (320 GB/s); prompt processing and long contexts are slower.

Measured speed on the Tesla T4

No community reports for this pair yet.

Report your own speed →

Run Arete-27B on the Tesla T4

Using Q3_K_M, the best fit for Tesla T4.

Ollama · Q3_K_M straight from Hugging Facedocs
ollama run hf.co/bartowski/NousResearch_Arete-27B-GGUF:Q3_K_M
llama.cpp · OpenAI-compatible server with a web UIdocs
llama-server -hf bartowski/NousResearch_Arete-27B-GGUF:Q3_K_M

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@Q3_K_M

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 Tesla T4 comfortably

All that fit →
ModelParamsBest quantMemorySpeedLicence
20.9BA4.2BQ8_0≈13 GBFits in memory~40 tok/sCommercial OK
14.8BQ6_K≈14 GBFits in memory~14 tok/sCommercial OK
microsoft logophi-4microsoft
14.7BQ6_K≈14 GBFits in memory~15 tok/sCommercial OK
14.8BQ6_K≈14 GBFits in memory~14 tok/sCommercial OK
8.2BQ8_0≈10 GBFits in memory~20 tok/sCommercial OK
8BQ8_0≈9.8 GBFits in memory~21 tok/sConditions

Other GPUs that run Arete-27B

Frequently asked questions

Can the Tesla T4 run Arete-27B?
Just about. At Q3_K_M it needs about 15 GB of the 16 GB available, so keep the context short, at roughly 13 tokens/s.
How fast does Arete-27B run on the Tesla T4?
An estimate from memory bandwidth gives about 13 tokens/s at Q3_K_M.
Which quantisation of Arete-27B should I use on the Tesla T4?
Q3_K_M: the largest one that fits. Expected quality: noticeable loss 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.