Zephyr 7B Beta Marlin is an open-source language model by RedHatAI. Features: 7b LLM, VRAM: 4.1GB, Context: 32K, Quantized, HF Score: 59.1, LLM Explorer Score: 0.25, ELO: 1053, Arc: 62, HellaSwag: 84.5, MMLU: 61.1, TruthfulQA: 57.4, WinoGrande: 78.1, GSM8K: 11.4.
Arxiv:2210.17323 4-bit Autotrain compatible Base model:huggingfaceh4/zephy... Base model:quantized:huggingfa... Conversational Endpoints compatible Gptq Int4 Marlin Mistral Nm-vllm Quantized Region:us Safetensors
Model Card on HF 🤗: https://huggingface.co/RedHatAI/zephyr-7b-beta-marlin
Zephyr 7B Beta Marlin Benchmarks
LMSys ELO: 1053 vs 1272 (so35)-17.2%
ARC: 62.03 vs 96.7 (so35)-35.9%
HellaSwag: 84.53 vs 95.3 (gpt4)-11.3%
MMLU: 61.06 vs 88.3 (so35)-30.8%
TruthfulQA: 57.44 vs 59 (gpt4)-2.6%
WinoGrande: 78.06 vs 87.5 (gpt4)-10.8%
GSM8K: 11.37 vs 96.4 (so35)-88.2%
LLME Score: 0.25348
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)
Zephyr 7B Beta Marlin Parameters and Internals
| LLM Name | Zephyr 7B Beta Marlin |
| Repository 🤗 | https://huggingface.co/RedHatAI/zephyr-7b-beta-marlin |
| Base Model(s) | |
| Model Size | 7b |
| Required VRAM | 4.1 GB |
| Updated | 2025-11-07 |
| Maintainer | RedHatAI |
| Model Type | mistral |
| Model Files | |
| GPTQ Quantization | Yes |
| Quantization Type | gptq |
| Model Architecture | MistralForCausalLM |
| Context Length | 32768 |
| Model Max Length | 32768 |
| Transformers Version | 4.37.2 |
| Tokenizer Class | LlamaTokenizer |
| Padding Token | </s> |
| Vocabulary Size | 32000 |
| Torch Data Type | float16 |
Best Alternatives to Zephyr 7B Beta Marlin
Best Alternatives |
Context / RAM |
Downloads |
Likes |
|---|---|---|---|
| ...enHermes 2.5 Mistral 7B Marlin | 32K / 4.1 GB | 584 | 2 |
| Mistral 7B Instruct V0.3 GPTQ | 32K / 4.2 GB | 62824 | 1 |
| Mistral 7B Instruct V0.2 GPTQ | 32K / 4.2 GB | 47360 | 54 |
| ...ral 7B Instruct V0.3 GPTQ 4bit | 32K / 4.2 GB | 2946 | 18 |
| ...ral 7B Instruct V0.3 GPTQ 4bit | 32K / 4.2 GB | 2095 | 23 |
| ...istral 7B Pruned50 GPTQ Marlin | 32K / 4 GB | 6 | 0 |
| Mistral 7B Unsloth Gptq 8bit | 32K / 7.7 GB | 7 | 0 |
| ...phyr 7B Beta Assistant V1 Gptq | 32K / 4.2 GB | 1 | 1 |
| ...l Neural Chat 7B V3.8 Bit Gptq | 32K / 7.7 GB | 7 | 0 |
| ...lai Mistral 7B V0.1 8 Bit Gptq | 32K / 7.7 GB | 8 | 0 |
Note: green Score (e.g. "73.2") means that the model is better than RedHatAI/zephyr-7b-beta-marlin .
Rank the Zephyr 7B Beta Marlin 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? 51648 in total.
Original data from HuggingFace, OpenCompass and various public git repos.
Check out Ag3ntum — our secure, self-hosted AI agent for server management.
Release v20260327b