What are the hardware requirements for Fiction.live Kimiko V2 70B GGML?

Fiction.live Kimiko V2 70B GGML requires approximately 28.6 GB of VRAM. Quantized variants may run on less VRAM; see the Quantized Models section on this page.

Who developed Fiction.live Kimiko V2 70B GGML and how large is it?

Fiction.live Kimiko V2 70B GGML is developed by TheBloke, a model with 70b 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 Fiction.live Kimiko V2 70B GGML?

Fiction.live Kimiko V2 70B GGML is hosted on Hugging Face and linked from this page. LLM Explorer also lists quantized variants and similar alternatives if available.

Fiction.live Kimiko V2 70B GGML by TheBloke — VRAM 28.6GB

Name: Fiction.live Kimiko V2 70B GGML
Author: TheBloke

Fiction.live Kimiko V2 70B GGML is an open-source language model by TheBloke. Features: 70b LLM, VRAM: 28.6GB, License: llama2, Quantized, LLM Explorer Score: 0.08.

Base model:chat-error/fiction.... Base model:finetune:chat-error... En Ggml Llama Quantized Region:us

Model Card on HF 🤗: https://huggingface.co/TheBloke/fiction.live-Kimiko-V2-70B-GGML

Fiction.live Kimiko V2 70B GGML Parameters and Internals

Model Type

large language model, text generation

Use Cases

Areas:

Research, Commercial, Roleplay

Applications:

Text Generation, Assistance, Roleplay

Primary Use Cases:

Normal and Erotic Roleplay

Limitations:

May have bias towards NSFW due to the large percentage of NSFW data in the training set

Additional Notes

Model is fine-tuned for both normal and erotic roleplay while also being capable as an assistant.

Training Details

Methodology:

Finetuning with LoRA

Context Length:

4090

Training Time:

10 hours

Hardware Used:

1 A100 for 10 hours on runpod

Input Output

Input Format:

text input following Fastchat/ShareGPT format

Accepted Modalities:

text

Output Format:

text

LLM Name	Fiction.live Kimiko V2 70B GGML
Repository 🤗	https://huggingface.co/TheBloke/fiction.live-Kimiko-V2-70B-GGML
Model Name	Fiction Live Kimiko V2 70B
Model Creator	nRuaif
Base Model(s)	nRuaif/fiction.live-Kimiko-V2-70B nRuaif/fiction.live-Kimiko-V2-70B
Model Size	70b
Required VRAM	28.6 GB
Updated	2026-08-06
Maintainer	TheBloke
Model Type	llama
Model Files	28.6 GB 36.1 GB 33.0 GB 29.7 GB 38.9 GB 43.2 GB 41.4 GB 38.9 GB 47.5 GB 48.8 GB 47.5 GB
Supported Languages	en
GGML Quantization	Yes
Quantization Type	ggml
Model Architecture	AutoModel
License	llama2

Rank the Fiction.live Kimiko V2 70B GGML 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 Fiction.live Kimiko V2 70B GGML

Best Alternatives	Context / RAM	Downloads	Likes
Synthia 70B V1.1 GGML	0K / 28.6 GB	3	4
...boros L2 70B 2.1 Creative GGML	0K / 28.6 GB	4	3
Llama 2 70B Chat GGML	0K / 28.6 GB	10	161
Lemur 70B Chat V1 GGML	0K / 29 GB	4	3
Model 007 70B GGML	0K / 28.6 GB	6	1
Airoboros L2 70B 2.1 GGML	0K / 28.6 GB	6	2
Nous Hermes Llama2 70B GGML	0K / 29 GB	4	13
Llama 2 70B Orca 200K GGML	0K / 28.6 GB	7	3
Genz 70B GGML	0K / 28.6 GB	6	3
Synthia 70B GGML	0K / 28.6 GB	3	2

Note: green Score (e.g. "73.2") means that the model is better than TheBloke/fiction.live-Kimiko-V2-70B-GGML.