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

DiscoLM 70B by DiscoResearch

 »  All LLMs  »  DiscoResearch  »  DiscoLM 70B   URL Share it on

  2662   

DiscoLM 70B is an open-source language model by DiscoResearch. Features: 70b LLM, VRAM: 138GB, Context: 8K, License: llama2, HF Score: 71.4, LLM Explorer Score: 0.14, Arc: 68.8, HellaSwag: 86.1, MMLU: 68.6, TruthfulQA: 57.6, WinoGrande: 83.6, GSM8K: 63.5.

  Autotrain compatible   Dataset:bjoernp/ultrachat de   Dataset:ldjnr/capybara   Dataset:leolm/german poems   Dataset:leolm/german songs   Dataset:leolm/openschnabeltier   Dataset:meta-math/metamathqa Dataset:migtissera/synthia-v1.... Dataset:open-orca/slimorca-ded...   Dataset:teknium/openhermes   Dataset:thudm/agentinstruct   De   Deutsch   Discoresearch   En   Endpoints compatible   Goliath   Llama   Llama2   Region:us   Safetensors   Sharded   Tensorflow
Model Card on HF 🤗: https://huggingface.co/DiscoResearch/DiscoLM-70b 

DiscoLM 70B Benchmarks

ARC: 68.77 vs 96.7 (so35)-28.9%
HellaSwag: 86.1 vs 95.3 (gpt4)-9.7%
MMLU: 68.58 vs 88.3 (so35)-22.3%
TruthfulQA: 57.64 vs 59 (gpt4)-2.3%
WinoGrande: 83.58 vs 87.5 (gpt4)-4.5%
GSM8K: 63.53 vs 96.4 (so35)-34.1%
LLME Score: 0.13586
nn.n% — How the model compares to the reference models: Anthropic Sonnet 3.5 ("so35"), GPT-4o ("gpt4o") or GPT-4 ("gpt4").
DiscoLM 70B (DiscoResearch/DiscoLM-70b)
🌟 Advertise your project 🚀

DiscoLM 70B Parameters and Internals

Model Type 
llama
Additional Notes 
Continued pretraining on German text improved multilingual capabilities.
Supported Languages 
en (partially improved), de (strengthened)
Training Details 
Data Sources:
Open-Orca/SlimOrca-Dedup, LeoLM/OpenSchnabeltier, teknium/openhermes, meta-math/MetaMathQA, bjoernp/ultrachat_de, migtissera/Synthia-v1.3, LeoLM/German_Songs, LeoLM/German_Poems, LDJnr/Capybara
Data Volume:
65 billion tokens of German text
Methodology:
Continued pretraining
Input Output 
Input Format:
ChatML format
Accepted Modalities:
text
Performance Tips:
Use the apply_chat_template() method for formatting messages.
LLM NameDiscoLM 70B
Repository 🤗https://huggingface.co/DiscoResearch/DiscoLM-70b 
Model CreatorDiscoResearch
Model Size70b
Required VRAM138 GB
Updated2025-09-23
MaintainerDiscoResearch
Model Typellama
Model Files  9.8 GB: 1-of-15   9.8 GB: 2-of-15   10.0 GB: 3-of-15   9.8 GB: 4-of-15   9.8 GB: 5-of-15   9.8 GB: 6-of-15   10.0 GB: 7-of-15   9.8 GB: 8-of-15   9.8 GB: 9-of-15   9.8 GB: 10-of-15   10.0 GB: 11-of-15   9.8 GB: 12-of-15   9.8 GB: 13-of-15   9.5 GB: 14-of-15   0.5 GB: 15-of-15
Supported Languagesen de
Model ArchitectureLlamaForCausalLM
Licensellama2
Context Length8192
Model Max Length8192
Transformers Version4.35.2
Tokenizer ClassLlamaTokenizer
Padding Token</s>
Vocabulary Size32032
Torch Data Typefloat16

Quantized Models of the DiscoLM 70B

Model
Likes
Downloads
VRAM
DiscoLM 70B GGUF429629 GB
DiscoLM 70B GPTQ4935 GB

Best Alternatives to DiscoLM 70B

Best Alternatives
Context / RAM
Downloads
Likes
... Chat 1048K Chinese Llama3 70B1024K / 141.9 GB90695
... Chat 1048K Chinese Llama3 70B1024K / 141.9 GB79144
... 3 70B Instruct Gradient 1048K1024K / 141.9 GB13122
Llama3 Function Calling 1048K1024K / 141.9 GB61
...a 3 70B Instruct Gradient 524K512K / 141.9 GB1023
...a 3 70B Instruct Gradient 262K256K / 141.9 GB2256
...ama 3 70B Arimas Story RP V2.0256K / 141.1 GB303
...ama 3 70B Arimas Story RP V1.6256K / 141.2 GB130
...ama 3 70B Arimas Story RP V1.5256K / 141.2 GB73
Yi 70B 200K RPMerge Franken195K / 142.4 GB31
Note: green Score (e.g. "73.2") means that the model is better than DiscoResearch/DiscoLM-70b.

Rank the DiscoLM 70B 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