Doradus-AI / text-generation updated 7 months ago

RnJ-1-Instruct-FP8

Name: RnJ-1-Instruct-FP8
Author: Doradus-AI

This is an FP8 quantized version of EssentialAI/RnJ-1-Instruct, created using llmcompressor (Neural Magic).

Params

8.8 B

Context

32,768

Downloads 30d

164 K

Likes

Commercial use: conditional · gemma Not gated SAFETENSORS 1 languages View on Hugging Face ↗

Download history

daily snapshots · 10 days

164 K164 K

Jul 28Jul 31Aug 3Aug 6

Can you run it?

Estimated VRAM at 8K context unless noted. Pick your hardware to see the verdict per quantization.

File	Quant	Size	Est. VRAM	Verdict on RTX 4090 · 24 GB
model.safetensors	f8_e4m3	12.0 GB	15.0 GB	✅ Runs comfortably
model.safetensors (bf16, full)	bf16 + 32K ctx	12.0 GB	19.0 GB	✅ Runs comfortably

Estimate: file size × 1.1 + KV cache at 8K + 0.5 GB overhead. Not a benchmark — how we calculate this.

Run it

copy-paste, exact tags checked against the Hub

~ · curl · api/v1

$ curl -s https://aimodelscomparison.com/api/v1/models/rnj-1-instruct-fp8
{
  "hf_id": "Doradus-AI/RnJ-1-Instruct-FP8",
  "params_b": 8.84,
  "context_length": 32768,
  "license": { "id": "gemma", "commercial": "conditional" },
  "downloads_30d": 163514,
  "vram_estimates": [
    { "quant": "f8_e4m3", "gb": 15.0 }
  ],
  "updated_at": "2026-07-28T18:08:56Z"
}

est. VRAM —on RTX 4090 · 24 GBJSON API →

Specifications

Architecture: Gemma3ForCausalLM
Parameters: 8.8 B
Tensor type: F8_E4M3
Context length: 32,768
Vocabulary: 128,256
Layers / heads: 32 / 32
Licence: gemma
First seen on the Hub: 2025-12-07
Training datasets: undisclosed
Added to our catalog: 2026-07-28

Compare with

Qwen3-8B 16.3 M Qwen2.5-7B-Instruct 12.1 M Qwen3.6-35B-A3B-NVFP4 11.3 M Llama-3.1-8B-Instruct 8.0 M Qwen2.5-7B-Instruct-AWQ 4.6 M granite-4.1-8b 4.2 M

Appears in

Best open coding models Commercial-use text models Models that fit in 12 GB