What are the hardware requirements for Flint 40B A2.5B MLX GPTQ MXFP4?

Flint 40B A2.5B MLX GPTQ MXFP4 requires approximately 21.3 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 40B A2.5B MLX GPTQ MXFP4 and how large is it?

Flint 40B A2.5B MLX GPTQ MXFP4 is developed by RESMP-DEV, a model with 40b 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 40B A2.5B MLX GPTQ MXFP4?

Flint 40B A2.5B MLX GPTQ MXFP4 is hosted on Hugging Face and linked from this page. LLM Explorer also lists quantized variants and similar alternatives if available.

Flint 40B A2.5B MLX GPTQ MXFP4 by RESMP-DEV — VRAM 21.3GB, 256K context

Name: Flint 40B A2.5B MLX GPTQ MXFP4
Author: RESMP-DEV

Flint 40B A2.5B MLX GPTQ MXFP4 is an open-source language model by RESMP-DEV. Features: 40b LLM, VRAM: 21.3GB, Context: 256K, License: apache-2.0, Quantized, LLM Explorer Score: 0.35.

4-bit Apple-silicon Base model:quantized:recursive... Base model:recursivesi/flint Conversational Flint Gptq Mlx Mlx-lm Moe Mxfp4 Quantized Region:us Safetensors Sharded Tensorflow

Model Card on HF 🤗: https://huggingface.co/RESMP-DEV/Flint-40B-A2.5B-MLX-GPTQ-MXFP4

Flint 40B A2.5B MLX GPTQ MXFP4 Parameters and Internals

LLM Name	Flint 40B A2.5B MLX GPTQ MXFP4
Repository 🤗	https://huggingface.co/RESMP-DEV/Flint-40B-A2.5B-MLX-GPTQ-MXFP4
Base Model(s)	Flint RecursiveSI/flint
Model Size	40b
Required VRAM	21.3 GB
Updated	2026-10-26
Maintainer	RESMP-DEV
Model Type	flint
Model Files	1.0 GB: 1-of-25 1.0 GB: 2-of-25 1.0 GB: 3-of-25 1.0 GB: 4-of-25 1.0 GB: 5-of-25 0.1 GB: 6-of-25 1.0 GB: 7-of-25 1.0 GB: 8-of-25 1.0 GB: 9-of-25 1.0 GB: 10-of-25 1.0 GB: 11-of-25 0.1 GB: 12-of-25 1.0 GB: 13-of-25 1.0 GB: 14-of-25 1.0 GB: 15-of-25 1.0 GB: 16-of-25 1.0 GB: 17-of-25 0.1 GB: 18-of-25 1.0 GB: 19-of-25 1.0 GB: 20-of-25 1.0 GB: 21-of-25 1.0 GB: 22-of-25 1.0 GB: 23-of-25 0.1 GB: 24-of-25 0.9 GB: 25-of-25
GPTQ Quantization	Yes
Quantization Type	gptq
Model Architecture	FlintForCausalLM
License	apache-2.0
Context Length	262144
Model Max Length	262144
Transformers Version	5.12.1
Tokenizer Class	TokenizersBackend
Vocabulary Size	150144
Torch Data Type	bfloat16

Rank the Flint 40B A2.5B MLX GPTQ MXFP4 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 40B A2.5B MLX GPTQ MXFP4

Best Alternatives	Context / RAM	Downloads	Likes
Flint 40B A2.5B Abliterated	256K / 79.6 GB	180	6
Flint 40B A2.5B JANG 6M	256K / 32.8 GB	305	1