RadixArk / text-generation updated 3 weeks ago

Qwen3.8-27B-DSpark

A DSpark speculative-decoding draft model for Qwen3.8-27B target models, trained with SpecForge and served with SGLang.

Params
1.9 B
Context
262,144
Downloads 30d
329 K
Likes
83
Licence: other Not gated SAFETENSORS View on Hugging Face ↗

Download history

daily snapshots · 20 days
416 K199 K
Sep 2Sep 8Sep 15Sep 21

Can you run it?

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

FileQuantSizeEst. VRAMVerdict on RTX 4090 · 24 GB
model.safetensors bf16 3.7 GB 4.9 GB ✅ Runs comfortably
model.safetensors (bf16, full) bf16 + 256K ctx 3.7 GB 13.5 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/qwen3-8-27b-dspark
{
  "hf_id": "RadixArk/Qwen3.8-27B-DSpark",
  "params_b": 1.86,
  "context_length": 262144,
  "license": { "id": "other", "commercial": "unknown" },
  "downloads_30d": 328736,
  "vram_estimates": [
    { "quant": "bf16", "gb": 4.9 }
  ],
  "updated_at": "2026-09-02T01:00:33Z"
}
est. VRAM —on RTX 4090 · 24 GBJSON API →

Specifications

Architecture
DSparkDraftModel
Parameters
1.9 B
Tensor type
BF16
Context length
262,144
Vocabulary
248,320
Layers / heads
5 / 32
Licence
other
First seen on the Hub
2026-08-14
Base model
Qwen3.8-27B-NVFP4
Training datasets
undisclosed
Added to our catalog
2026-09-02

Family

Base model and the most-downloaded derivatives in the catalog.

Compare with any text-generation model