What are the hardware requirements for Math SFT V4 4ksteps?

Math SFT V4 4ksteps requires approximately 1 GB of VRAM and supports a context window of 32K tokens. Quantized variants may run on less VRAM; see the Quantized Models section on this page.

Who developed Math SFT V4 4ksteps and how large is it?

Math SFT V4 4ksteps is developed by GabrielMM, a model with 494m 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 Math SFT V4 4ksteps?

Math SFT V4 4ksteps is hosted on Hugging Face and linked from this page. LLM Explorer also lists quantized variants and similar alternatives if available.

Math SFT V4 4ksteps by GabrielMM — VRAM 1GB, 32K context

Name: Math SFT V4 4ksteps
Author: GabrielMM

Math SFT V4 4ksteps is an open-source language model by GabrielMM. Features: 494m LLM, VRAM: 1GB, Context: 32K, LLM Explorer Score: 0.18.

Arxiv:1910.09700 Conversational Endpoints compatible Qwen2 Region:us Safetensors

Model Card on HF 🤗: https://huggingface.co/GabrielMM/Math_SFT_v4_4ksteps

Math SFT V4 4ksteps Parameters and Internals

LLM Name	Math SFT V4 4ksteps
Repository 🤗	https://huggingface.co/GabrielMM/Math_SFT_v4_4ksteps
Model Size	494m
Required VRAM	1 GB
Updated	2026-07-23
Maintainer	GabrielMM
Model Type	qwen2
Model Files	1.0 GB
Model Architecture	Qwen2ForCausalLM
Context Length	32768
Model Max Length	32768
Transformers Version	4.52.4
Tokenizer Class	Qwen2Tokenizer
Padding Token	<\|endoftext\|>
Vocabulary Size	151936
Torch Data Type	bfloat16
Errors	replace

Rank the Math SFT V4 4ksteps 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 Math SFT V4 4ksteps

Best Alternatives	Context / RAM	Downloads
Qwen2 Sft Warmup	128K / 0 GB	6
Qwen2 0.5Bchp 690K	128K / 2 GB	5
Qwen2 0.5Bchp 570K	128K / 2 GB	13
Qwen2 0.5Bchp 15K	128K / 2 GB	10
Qwen2 0.5Bchp 300000	128K / 2 GB	8
Qwen2 0.5Bchp 20000	128K / 2 GB	12
Svig Tiny Step 3.7K	128K / 2 GB	16
...en2 Emotions Without Reasoning	128K / 2 GB	11
Tinyqwen	128K / 1 GB	9
Luwa 01	32K / 2 GB	542

Note: green Score (e.g. "73.2") means that the model is better than GabrielMM/Math_SFT_v4_4ksteps.