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

Magnum V2 72B by anthracite-org

 »  All LLMs  »  anthracite-org  »  Magnum V2 72B   URL Share it on

  1903   
Base model:finetune:qwen/qwen2... Base model:qwen/qwen2-72b-inst...   Chat   Conversational Dataset:anthracite-org/kalo-op... Dataset:anthracite-org/nopm cl... Dataset:doctor-shotgun/c2-sthe...   De   En   Es   Fr   Instruct   It   Ja   Model-index   Pt   Qwen2   Region:us   Ru   Safetensors   Sharded   Tensorflow   Zh
Model Card on HF 🤗: https://huggingface.co/anthracite-org/magnum-v2-72b 

Magnum V2 72B Benchmarks

MMLU Pro: 49.51
GPQA: 18.12
MUSR: 14.18
BBH: 57.85
IFEval: 75.6 vs 88 (so35)-14.1%
MATH Lvl 5: 35.42
LLME Score: 0.36736
nn.n% — How the model compares to the reference models: Anthropic Sonnet 3.5 ("so35"), GPT-4o ("gpt4o") or GPT-4 ("gpt4").
Magnum V2 72B (anthracite-org/magnum-v2-72b)

Magnum V2 72B Parameters and Internals

Model Type 
text-generation
Additional Notes 
This is the seventh model in a series designed to replicate the prose quality of the Claude 3 models, specifically Sonnet and Opus.
Supported Languages 
en (Full), fr (Full), de (Full), es (Full), it (Full), pt (Full), ru (Full), zh (Full), ja (Full)
Training Details 
Data Sources:
anthracite-org/Stheno-Data-Filtered, anthracite-org/kalo-opus-instruct-22k-no-refusal, anthracite-org/nopm_claude_writing_fixed
Methodology:
Instruct tuned with ChatML formatting
Context Length:
16000
Hardware Used:
8x AMD Instinct MI300X Accelerators
Input Output 
Input Format:
ChatML format
Accepted Modalities:
text
LLM NameMagnum V2 72B
Repository 🤗https://huggingface.co/anthracite-org/magnum-v2-72b 
Base Model(s)  Qwen2 72B Instruct   Qwen/Qwen2-72B-Instruct
Model Size72b
Required VRAM146 GB
Updated2025-02-22
Maintaineranthracite-org
Model Typeqwen2
Instruction-BasedYes
Model Files  4.5 GB: 1-of-31   5.0 GB: 2-of-31   4.8 GB: 3-of-31   4.8 GB: 4-of-31   4.8 GB: 5-of-31   5.0 GB: 6-of-31   4.8 GB: 7-of-31   4.8 GB: 8-of-31   4.8 GB: 9-of-31   5.0 GB: 10-of-31   4.8 GB: 11-of-31   4.8 GB: 12-of-31   4.8 GB: 13-of-31   5.0 GB: 14-of-31   4.8 GB: 15-of-31   4.8 GB: 16-of-31   4.8 GB: 17-of-31   5.0 GB: 18-of-31   4.8 GB: 19-of-31   4.8 GB: 20-of-31   4.8 GB: 21-of-31   5.0 GB: 22-of-31   4.8 GB: 23-of-31   4.8 GB: 24-of-31   4.8 GB: 25-of-31   5.0 GB: 26-of-31   4.8 GB: 27-of-31   4.8 GB: 28-of-31   4.8 GB: 29-of-31   3.2 GB: 30-of-31   2.5 GB: 31-of-31
Supported Languagesen fr de es it pt ru zh ja
Model ArchitectureQwen2ForCausalLM
Licenseother
Context Length32768
Model Max Length32768
Transformers Version4.43.4
Tokenizer ClassQwen2Tokenizer
Padding Token<|endoftext|>
Vocabulary Size152064
Torch Data Typebfloat16
Errorsreplace

Best Alternatives to Magnum V2 72B

Best Alternatives
Context / RAM
Downloads
Likes
EVA Qwen2.5 72B V0.2128K / 146 GB239416
AceInstruct 72B128K / 146 GB18714
...n2.5 72B 2x Instruct TIES V1.0128K / 146.1 GB80
72B Qwen2.5 Kunou V1128K / 146 GB86021
...A Abliterated TIES Qwen2.5 72B128K / 146.1 GB520
EVA Qwen2.5 72B V0.1128K / 146 GB86115
EVA Qwen2.5 72B V0.0128K / 146 GB1035
Qwen2.5 72B Instruct32K / 145.5 GB354013741
Athene V2 Chat32K / 146 GB9149280
Qwen2 72B Instruct32K / 145.5 GB47866706
Note: green Score (e.g. "73.2") means that the model is better than anthracite-org/magnum-v2-72b.

Rank the Magnum V2 72B 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? 43470 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.
Release v20241227
  Support LLM Explorer