What are the hardware requirements for 34B Beta?

34B Beta requires approximately 69.2 GB of VRAM and supports a context window of 195K tokens. Quantized variants may run on less VRAM; see the Quantized Models section on this page.

Who developed 34B Beta and how large is it?

34B Beta is developed by CausalLM, a model with 34b parameters. The model is published as open weights on Hugging Face and indexed on LLM Explorer with full benchmark history.

How does 34B Beta perform on standard benchmarks?

34B Beta has the following published scores: MMLU 85.6. Compare against reference models on this page or on the LLM Explorer leaderboards.

34B Beta by CausalLM — VRAM 69.2GB, 195K context

Name: 34B Beta
Rating: 4 (26 reviews)
Author: CausalLM

34B Beta is an open-source language model by CausalLM. Features: 34b LLM, VRAM: 69.2GB, Context: 195K, License: gpl-3.0, LLM Explorer Score: 0.21, Arc: 70.6, HellaSwag: 84.2, MMLU: 85.6, GSM8K: 58.2.

Conversational Endpoints compatible Llama Region:us Safetensors Sharded Tensorflow

Model Card on HF 🤗: https://huggingface.co/CausalLM/34b-beta

34B Beta Parameters and Internals

Model Type

chat

Additional Notes

No repetition penalty.

Input Output

Input Format:

chatml

Performance Tips:

Use q8_0 quantization with llama.cpp for better performance than bf16 vllm for this model.

Release Notes

Version:

Next version update

Notes:

Issues with model weights in terms of precision will be fixed. Avoid using accelerated inference frameworks like VLLM temporarily. Use Transformers for inference.

LLM Name	34B Beta
Repository 🤗	https://huggingface.co/CausalLM/34b-beta
Model Size	34b
Required VRAM	69.2 GB
Updated	2026-09-18
Maintainer	CausalLM
Model Type	llama
Model Files	4.8 GB: 1-of-15 4.8 GB: 2-of-15 5.0 GB: 3-of-15 4.8 GB: 4-of-15 4.8 GB: 5-of-15 5.0 GB: 6-of-15 4.8 GB: 7-of-15 4.8 GB: 8-of-15 5.0 GB: 9-of-15 4.8 GB: 10-of-15 4.8 GB: 11-of-15 5.0 GB: 12-of-15 4.8 GB: 13-of-15 4.8 GB: 14-of-15 1.2 GB: 15-of-15
Model Architecture	LlamaForCausalLM
License	gpl-3.0
Context Length	200000
Model Max Length	200000
Transformers Version	4.37.2
Tokenizer Class	LlamaTokenizer
Padding Token	<unk>
Vocabulary Size	64000
Torch Data Type	bfloat16

Rank the 34B Beta capabilities

Have you tried this model? Rate its performance — this feedback helps the ML community find the right model for their needs.

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

Quantized Models of the 34B Beta

Model	Likes	Downloads	VRAM
34B Beta GGUF	4	146	12 GB

Best Alternatives to 34B Beta

Best Alternatives	Context / RAM	Downloads	Likes
Bagel Hermes 34B Slerp	195K / 68.9 GB	8375	1
Smaug 34B V0.1	195K / 69.2 GB	8082	64
Casual Magnum 34B	195K / 68.8 GB	4	1
Yi 34B 200K	195K / 68.9 GB	9133	320
Bagel 34B V0.2	195K / 68.7 GB	259	41
Yi 34B 200K AEZAKMI V2	195K / 69.2 GB	57	12
Faro Yi 34B	195K / 69.2 GB	8208	6
Bagel DPO 34B V0.5	195K / 68.7 GB	8040	17
Smaug 34B V0.1 ExPO	195K / 69.2 GB	7812	0
Luminex 34B V0.1	195K / 68.9 GB	8566	8

Note: green Score (e.g. "73.2") means that the model is better than CausalLM/34b-beta.