Model library / TokenRhythm
NeoHorse-1-9B is a 9 billion-parameter causal language model fine-tuned for text-based agent tasks, tool use, coding and instruction following.
Parameters in total
Context length in tokens, as published
Weights at the released precision
Licence declared on the repository
Overview
NeoHorse-1-9B is a 9 billion-parameter causal language model derived from Qwen3.5-9B and fine-tuned by TokenRhythm. The publisher describes it as an initial prototype aimed at recursive self-improvement, employing a routing harness that records tool interactions and informs future training mixtures. It targets text-only inference for agentic applications, tool use and coding assistance.
In practice the model occupies roughly the size of a typical 9 B parameter transformer, uses bfloat16 weights and runs with a context window of 262 144 tokens. It is released under the Apache-2.0 licence, has no mixture-of-experts layers, and is configured as Qwen3_5ForCausalLM for qwen3_5_text tasks. These characteristics define its resource requirements and licensing terms for deployment.
Specification
| Repository | TokenRhythm/NeoHorse-1-9B |
| Publisher | TokenRhythm |
| Published | 5 September 2026 |
| Architecture | Qwen3_5ForCausalLM |
| Model type | qwen3_5_text |
| Parameters | 8,953,803,264 |
| Layers | 32 |
| Hidden size | 4,096 |
| Attention heads | 16, 4 key/value heads |
| Context length | 262,144 tokens |
| Vocabulary | 248,320 tokens |
| Weight format | bfloat16 |
| Licence | apache-2.0 |
| Base model | Qwen/Qwen3.5-9B |
Memory
Parameter count multiplied by the bytes each format uses. The key-value cache comes on top.
| Precision | Weights | Notes |
|---|---|---|
| bf16 | 16.7 GiB | Full precision as released |
| fp8 | 8.3 GiB | 8-bit, near-lossless on most models |
| int4 | 4.2 GiB | 4-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
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.
| GPU | Cards needed | Free for cache | Context it holds |
|---|---|---|---|
| RTX 4090 24GB | 1 | 2.4 GiB | about 19,843 tokens |
| RTX 5090 32GB | 1 | 9.6 GiB | about 78,825 tokens |
| L40S 48GB | 1 | 24.0 GiB | about 196,790 tokens |
| A100 80GB | 1 | 52.8 GiB | the full 262,144 tokens |
| H100 80GB | 1 | 52.8 GiB | the full 262,144 tokens |
| RTX PRO 6000 96GB | 1 | 67.2 GiB | the full 262,144 tokens |
| H200 141GB | 1 | 107.7 GiB | the full 262,144 tokens |
Serving
A vLLM launch line for the shape above. Check the model card for a runtime the publisher recommends.
vllm serve TokenRhythm/NeoHorse-1-9B \ --max-model-len 16384 \ --gpu-memory-utilization 0.90
Questions
The weights take about 17 GiB at the released precision. That fits on one RTX 4090 24GB, which leaves roughly 2 GiB for the key-value cache.
The repository declares apache-2.0. Read the licence text before commercial use: the name alone does not tell you what is allowed.
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.
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
NeoHorse-1-4B is a 4.2B open-weight model published by TokenRhythm on Hugging Face. It uses the Qwen3_5ForCausalLM architecture w…
ZGCM-1 is a 7.4 billion-parameter dense language model designed for mathematical reasoning and tool-assisted search.
LLaDA2.2-mini is a 16.3 billion-parameter MoE diffusion language model with 128 k token context, supporting Levenshtein editing.
K2-Horizon-7B is a 9-billion-parameter decoder-only language model with a 524k token context window, released under Apache-2.0.
Hosting
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.