Decision models ·

Prometheus 2 8x7B

Task
Judging and grading answers
Returns
Written feedback, then a score
Precision tier
74 / 100
Size
47B (13B active)
Max input
32K tokens
License
apache-2.0
Released
2024-05

What it decides

Returns
Feedback text, then a 1 to 5 score or A/B preference against your rubric.
Checks
Rubric scoring and pairwise ranking
Languages
English
Precision
74 / 100 · Highest agreement with human and GPT-4 judgments among open evaluators at release (paper).
Clarity
Written feedback, then a score (9 of 15 clarity points) · writes about 150 tokens before the answer, so it is slower but explains itself.
How it runs
Writes a short answer · a small language model, so it runs in the same apps as LLMs.
License
apache-2.0 (Open (Apache, MIT))

Source: model card on Hugging Face (opens in a new tab) · How decision models are rated

How to run it

Quantized files · GGUF

Works in LM Studio, llama.cpp, Jan and Ollama. No sign-in needed.

huggingface.co/mradermacher/prometheus-8x7b-v2.0-GGUF (opens in a new tab)

ollama run hf.co/mradermacher/prometheus-8x7b-v2.0-GGUF

Original weights · Hugging Face

Full-precision files from Prometheus Eval, for fine-tuning, vLLM, Transformers.

huggingface.co/prometheus-eval/prometheus-8x7b-v2.0 (opens in a new tab)

No sign-in needed.

Links checked 3 days ago against Hugging Face, Ollama and GitHub.

Apps that run Prometheus 2 8x7B

Our pick for each kind of computer, assuming the model fits in its memory. Pick your own GPU or Mac in the Model Finder for an exact answer.

Mac (Apple silicon)

LM Studio

Point-and-click: search, download and chat in one window.

Also good: llama.cpp, Ollama

Windows PC, NVIDIA GPU

LM Studio

Point-and-click: search, download and chat in one window.

Also good: llama.cpp, Ollama

Windows PC, AMD Radeon

Lemonade

AMD’s own app: llama.cpp tuned for Radeon on Vulkan and ROCm.

Also good: LM Studio, llama.cpp

Linux, NVIDIA GPU

LM Studio

Point-and-click: search, download and chat in one window.

Also good: llama.cpp, Ollama

Linux server, 80 GB+ GPUs

vLLM

The full model fits in GPU memory at 8-bit, so it can batch many requests: best for serving a team or an app.

Also good: LM Studio, SGLang

App Type Opens it as Systems Engine Start
LM Studio Desktop app GGUF macOSWindowsLinux llama.cpp, MLX Open in LM Studio
Ollama CLI + API GGUF from Hugging Face macOSWindowsLinux llama.cpp, MLX 
ollama run hf.co/mradermacher/prometheus-8x7b-v2.0-GGUF:Q4_K_M
llama.cpp CLI + API GGUF macOSWindowsLinux llama.cpp 
llama-server -hf mradermacher/prometheus-8x7b-v2.0-GGUF:Q4_K_M
Jan Desktop app GGUF macOSWindowsLinux llama.cpp, MLX Open in Jan
vLLM Server Original weights WSLWindowsLinux vLLM 
vllm serve prometheus-eval/prometheus-8x7b-v2.0
SGLang Server Original weights WSLWindowsLinux SGLang 
python3 -m sglang.launch_server --model-path prometheus-eval/prometheus-8x7b-v2.0
Docker Model Runner CLI + API GGUF macOSWindowsLinux llama.cpp, vLLM 
docker model run hf.co/mradermacher/prometheus-8x7b-v2.0-GGUF:Q4_K_M
Lemonade Desktop app GGUF macOSWindowsLinux llama.cpp, ONNX Runtime GenAI 
lemonade pull mradermacher/prometheus-8x7b-v2.0-GGUF:Q4_K_M
Unsloth Desktop Desktop app GGUF macOSWindowsLinux llama.cpp, MLX, stable-diffusion.cpp Open in Unsloth Desktop
Atomic Chat Desktop app GGUF macOSWindowsLinux llama.cpp, MLX-VLM Open in Atomic Chat
LocalAI Server GGUF macOSWSLWindowsLinux llama.cpp, rerankers, Transformers Get LocalAI
Sentence Transformers Library Original weights macOSWindowsLinux PyTorch Get Sentence Transformers

Which apps can open a model depends on the files published for it: GGUF for llama.cpp-based apps, MLX for Apple’s engine, the original weights for vLLM and SGLang. Compare every app.

Memory and speed

Each quantization on an NVIDIA GeForce RTX 4090, for one decision on a 1,024-token input. The best hardware table below covers other GPUs and Macs.

Format Memory needed Time per decision Decisions per second Quality Download
Q4_K_M 29.4 GB 13.6 s 0.1/s 70Great quality (A) · 70 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 51/100 (quality, fit, speed and memory use) sets the ranking. GGUF (opens in a new tab)
Q5_K_M 34.0 GB 22.2 s 0.0/s 71Great quality (A) · 71 of 100: how capable the model is at Q5_K_M, whatever it runs on. Match score 46/100 (quality, fit, speed and memory use) sets the ranking. GGUF (opens in a new tab)
Q6_K 39.2 GB 32.0 s 0.0/s 73Great quality (A) · 73 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 46/100 (quality, fit, speed and memory use) sets the ranking. GGUF (opens in a new tab)
Q8_0 50.4 GB — — 74Great quality (A) · 74 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 9/100 (quality, fit, speed and memory use) sets the ranking. GGUF (opens in a new tab)
F16 94.3 GB — — 74Great quality (A) · 74 of 100: how capable the model is at F16, whatever it runs on. Match score 9/100 (quality, fit, speed and memory use) sets the ranking. Original weights (opens in a new tab)

One request at a time. Servers that batch (Text Embeddings Inference, vLLM, Infinity) handle many more decisions per second.

Best hardware for Prometheus 2 8x7B

The least expensive GPU in our catalog that runs it without offloading is the Intel Arc Pro B70 (launch price $949) at Q4_K_M.

Quality Hardware Quant Memory Time per decision Runs
73Great quality (A) · 73 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 72/100 (quality, fit, speed and memory use) sets the ranking. NVIDIA B200NVIDIA B200 Q6_K 48.2 GB / 180 GB 687 ms Slow
73Great quality (A) · 73 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 72/100 (quality, fit, speed and memory use) sets the ranking. AMD Instinct MI350XAMD Instinct MI350X Q6_K 53.6 GB / 288 GB 819 ms Slow
73Great quality (A) · 73 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 71/100 (quality, fit, speed and memory use) sets the ranking. AMD Instinct MI325XAMD Instinct MI325X Q6_K 52.0 GB / 256 GB 1.0 s Slow
73Great quality (A) · 73 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 71/100 (quality, fit, speed and memory use) sets the ranking. NVIDIA H200 141GB SXMNVIDIA H200 141GB SXM Q6_K 46.2 GB / 141 GB 1.1 s Slow
73Great quality (A) · 73 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 71/100 (quality, fit, speed and memory use) sets the ranking. NVIDIA H200 NVLNVIDIA H200 NVL Q6_K 46.2 GB / 141 GB 1.1 s Slow
73Great quality (A) · 73 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 71/100 (quality, fit, speed and memory use) sets the ranking. AMD Instinct MI300XAMD Instinct MI300X Q6_K 48.8 GB / 192 GB 1.1 s Slow
73Great quality (A) · 73 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 71/100 (quality, fit, speed and memory use) sets the ranking. NVIDIA GH200 Grace Hopper (96 GB GPU)NVIDIA GH200 Grace Hopper (96 GB GPU) Q6_K 44.0 GB / 96.0 GB 1.3 s Slow
73Great quality (A) · 73 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 71/100 (quality, fit, speed and memory use) sets the ranking. NVIDIA H100 NVLNVIDIA H100 NVL Q6_K 43.9 GB / 94.0 GB 1.3 s Slow
71Great quality (A) · 71 of 100: how capable the model is at Q5_K_M, whatever it runs on. Match score 70/100 (quality, fit, speed and memory use) sets the ranking. NVIDIA H100 80GB SXMNVIDIA H100 80GB SXM Q5_K_M 38.0 GB / 80.0 GB 1.3 s Slow
71Great quality (A) · 71 of 100: how capable the model is at Q5_K_M, whatever it runs on. Match score 70/100 (quality, fit, speed and memory use) sets the ranking. NVIDIA H800 80GB SXMNVIDIA H800 80GB SXM Q5_K_M 38.0 GB / 80.0 GB 1.3 s Slow

Check it on your own hardware

Estimates come from published hardware specs and our open formulas, not from running the model on your machine. Real speed varies with runtime, drivers and settings. Numbers marked “Measured” come from a linked benchmark.