What are the hardware requirements for Flint MLX Serve 4bit?

Flint MLX Serve 4bit requires approximately 22.5 GB of VRAM and supports a context window of 256K tokens. Quantized variants may run on less VRAM; see the Quantized Models section on this page.

Who developed Flint MLX Serve 4bit and how large is it?

Flint MLX Serve 4bit is developed by pgmharikrishnan, a model with 39.8b 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 Flint MLX Serve 4bit?

Flint MLX Serve 4bit is hosted on Hugging Face and linked from this page. LLM Explorer also lists quantized variants and similar alternatives if available.

Flint MLX Serve 4bit by pgmharikrishnan — VRAM 22.5GB, 256K context

Name: Flint MLX Serve 4bit
Author: pgmharikrishnan

Flint MLX Serve 4bit is an open-source language model by pgmharikrishnan. Features: 39.8b LLM, VRAM: 22.5GB, Context: 256K, License: apache-2.0, Quantized, LLM Explorer Score: 0.35.

4-bit 4bit Apple-silicon Base model:quantized:recursive... Base model:recursivesi/flint Conversational Flint Mlx Mlx-serve Moe Quantized Region:us Safetensors Sharded Tensorflow

Model Card on HF 🤗: https://huggingface.co/pgmharikrishnan/Flint-MLX-Serve-4bit

Flint MLX Serve 4bit Parameters and Internals

LLM Name	Flint MLX Serve 4bit
Repository 🤗	https://huggingface.co/pgmharikrishnan/Flint-MLX-Serve-4bit
Base Model(s)	Flint RecursiveSI/flint
Model Size	39.8b
Required VRAM	22.5 GB
Updated	2026-10-25
Maintainer	pgmharikrishnan
Model Type	flint
Model Files	4.4 GB: 1-of-6 4.4 GB: 2-of-6 4.4 GB: 3-of-6 4.3 GB: 4-of-6 4.4 GB: 5-of-6 0.6 GB: 6-of-6
Quantization Type	4bit
Model Architecture	FlintForCausalLM
License	apache-2.0
Context Length	262144
Model Max Length	262144
Transformers Version	5.12.1
Tokenizer Class	PreTrainedTokenizerFast
Vocabulary Size	150144
Torch Data Type	bfloat16

Rank the Flint MLX Serve 4bit 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 Flint MLX Serve 4bit

Best Alternatives	Context / RAM	Downloads	Likes
Flint MLX Serve Mixed 4 8bit	256K / 23.4 GB	1230	3
Flint	256K / 80 GB	4795	754
Flint FP8 BF16	256K / 0.8 GB	46	1

Note: green Score (e.g. "73.2") means that the model is better than pgmharikrishnan/Flint-MLX-Serve-4bit.