Aiqre

Model library / TokenRhythm

NeoHorse-1-4B

NeoHorse-1-4B is a 4.2B open-weight model published by TokenRhythm on Hugging Face. It uses the Qwen3_5ForCausalLM architecture with 32 layers. The published configuration allows a context of 262,144 tokens. The weights are released under the apache-2.0 licence.

4.2B

Parameters in total

262,144

Context length in tokens, as published

8 GiB

Weights at the released precision

apache-2.0

Licence declared on the repository

Overview

What this model is.

NeoHorse-1-4B is a 4.2B open-weight model published by TokenRhythm on Hugging Face. It uses the Qwen3_5ForCausalLM architecture with 32 layers. The published configuration allows a context of 262,144 tokens. The weights are released under the apache-2.0 licence.

At the released precision the weights occupy about 8 GiB, before any key-value cache. That fits on a single RTX 4090 24GB. The tables below work the memory out for common cards.

Specification

The published shape.

RepositoryTokenRhythm/NeoHorse-1-4B
PublisherTokenRhythm
Published5 September 2026
ArchitectureQwen3_5ForCausalLM
Model typeqwen3_5_text
Parameters4,205,751,296
Layers32
Hidden size2,560
Attention heads16, 4 key/value heads
Context length262,144 tokens
Vocabulary248,320 tokens
Weight formatbfloat16
Licenceapache-2.0
Base modelQwen/Qwen3.5-4B

Memory

How much VRAM the weights need.

Parameter count multiplied by the bytes each format uses. The key-value cache comes on top.

PrecisionWeightsNotes
bf167.8 GiBFull precision as released
fp83.9 GiB8-bit, near-lossless on most models
int42.0 GiB4-bit, smallest footprint

The cache costs 128 KiB per token at 16-bit, so the full 262,144-token context of one request needs about 32.0 GiB. Concurrency multiplies that number, not the weights.

Hardware

Which card runs it.

At the released precision, with 90% of the card given to the server and 2.5 GiB kept for the runtime. The context column is what is left for the cache on that setup.

GPUCards neededFree for cacheContext it holds
RTX 4090 24GB 1 11.3 GiB about 92,292 tokens
RTX 5090 32GB 1 18.5 GiB about 151,274 tokens
L40S 48GB 1 32.9 GiB the full 262,144 tokens
A100 80GB 1 61.7 GiB the full 262,144 tokens
H100 80GB 1 61.7 GiB the full 262,144 tokens
RTX PRO 6000 96GB 1 76.1 GiB the full 262,144 tokens
H200 141GB 1 116.6 GiB the full 262,144 tokens

Serving

Running it yourself.

A vLLM launch line for the shape above. Check the model card for a runtime the publisher recommends.

vllm serve TokenRhythm/NeoHorse-1-4B \
  --max-model-len 90112 \
  --gpu-memory-utilization 0.90

Questions

The things people ask about this model.

How much GPU memory does NeoHorse-1-4B need?

The weights take about 8 GiB at the released precision. That fits on one RTX 4090 24GB, which leaves roughly 11 GiB for the key-value cache.

What licence does NeoHorse-1-4B use?

The repository declares apache-2.0. Read the licence text before commercial use: the name alone does not tell you what is allowed.

How long a context does it support?

The configuration allows 262,144 tokens. Whether the whole window is usable depends on the memory left for the cache, which the hardware table works out per card.

Can I use it through an API instead of hosting it?

This page is a reference, and Aiqre does not serve this model today. We deploy open-weight models on dedicated EU hardware on request, so if you want an OpenAI-compatible endpoint for it, email [email protected] with the model name and roughly what volume you expect.

Also in the library

Related models.

TokenRhythm/NeoHorse-1-9B

NeoHorse-1-9B is a 9 billion-parameter causal language model fine-tuned for text-based agent tasks, tool use, coding and instruct…

zgcagi/ZGCM-1-7B

ZGCM-1 is a 7.4 billion-parameter dense language model designed for mathematical reasoning and tool-assisted search.

openbmb/MiniCPM5-2B

MiniCPM5-2B is a 2-billion-parameter dense transformer aimed at on-device and resource-constrained deployments.

IFM/K2-Horizon-3.7B

K2-Horizon-3.7B is a 3.7 billion-parameter dense decoder-only model with a 512 K token context, released openly by IFM.

Hosting

Want this model on a dedicated EU GPU?

Tell us the model and roughly what volume you expect. We reply with a price and an OpenAI-compatible endpoint, with your prompts never stored.