4-bit 4bit Base model:deepseek-ai/deepsee... Base model:quantized:deepseek-... Conversational Llama Mlx Quantized Region:us Safetensors
Model Card on HF 🤗: https://huggingface.co/mlx-community/DeepSeek-Prover-V2-7B-4bit
DeepSeek Prover V2 7B 4bit Benchmarks
nn.n% — How the model compares to the reference models: Anthropic Sonnet 3.5 ("so35"), GPT-4o ("gpt4o") or GPT-4 ("gpt4").
What is the LLM Explorer Rank (Score)
DeepSeek Prover V2 7B 4bit Parameters and Internals
| LLM Name | DeepSeek Prover V2 7B 4bit |
| Repository 🤗 | https://huggingface.co/mlx-community/DeepSeek-Prover-V2-7B-4bit |
| Base Model(s) | |
| Model Size | 7b |
| Required VRAM | 3.9 GB |
| Updated | 2025-05-17 |
| Maintainer | mlx-community |
| Model Type | llama |
| Model Files | |
| Quantization Type | 4bit |
| Model Architecture | LlamaForCausalLM |
| Context Length | 65536 |
| Model Max Length | 65536 |
| Transformers Version | 4.46.3 |
| Tokenizer Class | LlamaTokenizerFast |
| Padding Token | <｜end▁of▁sentence｜> |
| Vocabulary Size | 102400 |
| Torch Data Type | bfloat16 |
Best Alternatives to DeepSeek Prover V2 7B 4bit
Best Alternatives |
Context / RAM |
Downloads |
Likes |
|---|---|---|---|
| Smaugv0.1 6.0bpw H6 EXL2 | 195K / 26.4 GB | 6 | 4 |
| Smaugv0.1 5.0bpw H6 EXL2 | 195K / 22.3 GB | 5 | 3 |
| Smaugv0.1 3.0bpw H6 EXL2 | 195K / 13.9 GB | 5 | 1 |
| Smaugv0.1 4.0bpw H6 EXL2 | 195K / 18 GB | 5 | 1 |
| Smaugv0.1 8.0bpw H8 EXL2 | 195K / 34.9 GB | 5 | 1 |
| Smaugv0.1 4.65bpw H6 EXL2 | 195K / 20.8 GB | 3 | 1 |
| Mistral 7B Openplatypus 1K | 32K / 29 GB | 7 | 0 |
| ...rnlm2 20B Llama 4.0bpw H6 EXL2 | 32K / 11 GB | 7 | 1 |
| Mistral 7B OpenOrca 1K | 32K / 29 GB | 16 | 3 |
| Mistral 7B A U0.5 B2 Ver0.4 | 32K / 14.4 GB | 1034 | 0 |
Rank the DeepSeek Prover V2 7B 4bit Capabilities
🆘 Have you tried this model? Rate its performance. This feedback would greatly assist ML community in identifying the most suitable model for their needs. Your contribution really does make a difference! 🌟
What open-source LLMs or SLMs are you in search of? 47402 in total.
Original data from HuggingFace, OpenCompass and various public git repos.
Release v20241227