LLM EXPLORER
 LLM Explorer: A Curated Large Language Model Directory and Analytics
Dark Theme
  Agents   Benchmarks   LLM  List    LLM Hosting    LLM Leaderboards    Blog    Newsfeed    Advertise

Dolphin 2.7 Mixtral 8x7b 4.0bpw H6 EXL2 by LoneStriker

 »  All LLMs  »  LoneStriker  »  Dolphin 2.7 Mixtral 8x7b 4.0bpw H6 EXL2   URL Share it on

  4   

Dolphin 2.7 Mixtral 8x7b 4.0bpw H6 EXL2 is an open-source language model by LoneStriker. Features: LLM, VRAM: 23.6GB, Context: 32K, License: apache-2.0, MoE, Quantized, Instruction-Based, LLM Explorer Score: 0.11.

  Conversational   Dataset:ehartford/dolphin Dataset:ehartford/dolphin-code... Dataset:ise-uiuc/magicoder-evo... Dataset:ise-uiuc/magicoder-oss... Dataset:jondurbin/airoboros-2....   Dataset:ldjnr/capybara   Dataset:teknium/openhermes   En   Endpoints compatible   Exl2   Instruct   Mixtral   Moe   Pytorch   Quantized   Region:us   Sharded   Tensorflow
Model Card on HF 🤗: https://huggingface.co/LoneStriker/dolphin-2.7-mixtral-8x7b-4.0bpw-h6-exl2 

Dolphin 2.7 Mixtral 8x7b 4.0bpw H6 EXL2 Benchmarks

LLME Score: 0.11287
nn.n% — How the model compares to the reference models: Anthropic Sonnet 3.5 ("so35"), GPT-4o ("gpt4o") or GPT-4 ("gpt4").
Dolphin 2.7 Mixtral 8x7b 4.0bpw H6 EXL2 (LoneStriker/dolphin-2.7-mixtral-8x7b-4.0bpw-h6-exl2)
🌟 Advertise your project 🚀

Dolphin 2.7 Mixtral 8x7b 4.0bpw H6 EXL2 Parameters and Internals

Model Type 
coding
Additional Notes 
You are advised to implement your own alignment layer before exposing the model as a service, as it is uncensored.
Supported Languages 
en (English)
Training Details 
Data Sources:
ehartford/dolphin, jondurbin/airoboros-2.2.1, ehartford/dolphin-coder, teknium/openhermes, ise-uiuc/Magicoder-OSS-Instruct-75K, ise-uiuc/Magicoder-Evol-Instruct-110K, LDJnr/Capybara
Methodology:
Retrained with Mixtral-specific fixes in transformers library and with gate layer unfrozen
Context Length:
16000
Training Time:
3 days to train 1.5 epochs
Hardware Used:
4x A100s
Input Output 
Input Format:
ChatML prompt format
Accepted Modalities:
text
Output Format:
Text
Performance Tips:
Encouragement needed in system prompt as model is not DPO tuned.
Release Notes 
Version:
2.7
Notes:
Retrained with Mixtral-specific fixes in transformers library, with gate layer unfrozen to fix performance issues.
Version:
2.6
Notes:
Fixed training configuration issue, improved quality, reintroduced samantha-based empathy data and replaced synthia and pure-dove with Capybara.
LLM NameDolphin 2.7 Mixtral 8x7b 4.0bpw H6 EXL2
Repository 🤗https://huggingface.co/LoneStriker/dolphin-2.7-mixtral-8x7b-4.0bpw-h6-exl2 
Required VRAM23.6 GB
Updated2026-03-31
MaintainerLoneStriker
Model Typemixtral
Instruction-BasedYes
Model Files  8.6 GB: 1-of-3   8.6 GB: 2-of-3   6.4 GB: 3-of-3
Supported Languagesen
Quantization Typeexl2
Model ArchitectureMixtralForCausalLM
Licenseapache-2.0
Context Length32768
Model Max Length32768
Transformers Version4.37.0.dev0
Tokenizer ClassLlamaTokenizer
Padding Token</s>
Vocabulary Size32002
Torch Data Typebfloat16

Best Alternatives to Dolphin 2.7 Mixtral 8x7b 4.0bpw H6 EXL2

Best Alternatives
Context / RAM
Downloads
Likes
...M 2 8x22B Beige 5.0bpw H6 EXL264K / 88.5 GB110
...M 2 8x22B Beige 2.4bpw H6 EXL264K / 42.7 GB60
...M 2 8x22B Beige 3.0bpw H6 EXL264K / 53.2 GB60
...M 2 8x22B Beige 4.0bpw H6 EXL264K / 70.8 GB50
...B Instruct V0.1 8.0bpw H8 EXL264K / 120.2 GB31
...8x22b Instruct Oh EXL2 2.25bpw64K / 40.1 GB21
...eryTour V2 8x7B 4.5bpw H6 EXL232K / 26.5 GB42
...it MoE 2bitgs8 Metaoffload HQQ32K / 24.1 GB2819
... 4bit MoE 3bit Metaoffload HQQ32K / 22.4 GB213
...hin 2.7 Mixtral 8x7b 8bpw EXL232K / 46.8 GB12
Note: green Score (e.g. "73.2") means that the model is better than LoneStriker/dolphin-2.7-mixtral-8x7b-4.0bpw-h6-exl2.

Rank the Dolphin 2.7 Mixtral 8x7b 4.0bpw H6 EXL2 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! 🌟

Instruction Following and Task Automation  
Factuality and Completeness of Knowledge  
Censorship and Alignment  
Data Analysis and Insight Generation  
Text Generation  
Text Summarization and Feature Extraction  
Code Generation  
Multi-Language Support and Translation  

What open-source LLMs or SLMs are you in search of? 52392 in total.

Email us: info@extractum.io. Our Privacy Policy  | Terms and Conditions  | Suggest an improvement.
Our Social Media →  
Original data from HuggingFace, OpenCompass and various public git repos.
Check out Ag3ntum — our secure, self-hosted AI agent for server management.
Release v20260328a
  Support LLM Explorer