What are the hardware requirements for RWKV V4 Raven 14B One State?

RWKV V4 Raven 14B One State requires approximately 28.3 GB of VRAM. Quantized variants may run on less VRAM; see the Quantized Models section on this page.

Who developed RWKV V4 Raven 14B One State and how large is it?

RWKV V4 Raven 14B One State is developed by xiaol, a model with 14b parameters. The model is published as open weights on Hugging Face and indexed on LLM Explorer with full benchmark history.

Where can I download or evaluate RWKV V4 Raven 14B One State?

RWKV V4 Raven 14B One State is hosted on Hugging Face and linked from this page. LLM Explorer also lists quantized variants and similar alternatives if available.

RWKV V4 Raven 14B One State by xiaol — VRAM 28.3GB

Name: RWKV V4 Raven 14B One State
Author: xiaol

RWKV V4 Raven 14B One State is an open-source language model by xiaol. Features: 14b LLM, VRAM: 28.3GB, License: apache-2.0, LLM Explorer Score: 0.09.

Dataset:baai/coig-pc Dataset:cais/mmlu Dataset:ehartford/dolphin Dataset:natural questions Dataset:norquinal/claude multi... Dataset:openleecher/teatime Dataset:pygmalionai/pippa Endpoints compatible Pytorch Region:us Rwkv Safetensors

Model Card on HF 🤗: https://huggingface.co/xiaol/RWKV-v4-raven-14B-one-state

RWKV V4 Raven 14B One State Parameters and Internals

Model Type

Role play, Instruction following, OpenLLM leaderboard

Additional Notes

The model has been finetuned on instruction datasets and can perform role play.

Training Details

Data Sources:

PygmalionAI/PIPPA, Norquinal/claude_multiround_chat_30k, ehartford/dolphin, cais/mmlu, OpenLeecher/Teatime, BAAI/COIG-PC, natural_questions

LLM Name	RWKV V4 Raven 14B One State
Repository 🤗	https://huggingface.co/xiaol/RWKV-v4-raven-14B-one-state
Model Size	14b
Required VRAM	28.3 GB
Updated	2026-07-13
Maintainer	xiaol
Model Type	rwkv
Model Files	28.3 GB 28.3 GB
Model Architecture	AutoModel
License	apache-2.0
Transformers Version	4.33.1
Tokenizer Class	PreTrainedTokenizerFast
Vocabulary Size	50277

Rank the RWKV V4 Raven 14B One State 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

Best Alternatives to RWKV V4 Raven 14B One State

Best Alternatives	Context / RAM	Downloads	Likes
....5 14B Instruct Abliterated V2	0K / 0.6 GB	0	3
Ct2fast Docsgpt 14B	0K / 13 GB	3	1
Phi 4 14B 1M RRP V1 Lora	0K / 0.3 GB	0	1
....5 Coder 14B Instruct F16 GGUF	0K / 29.5 GB	180	2
LWM 14B Text Chat 1M GGUF	0K / 2.5 GB	584	0
CausalLM 14B GGUF	0K / 8.2 GB	4262	196

Note: green Score (e.g. "73.2") means that the model is better than xiaol/RWKV-v4-raven-14B-one-state.