flint-GGUF
See Unsloth Dynamic 3.0 GGUFs for our quantization benchmarks. See further below for quant benchmarks. See below for Flint running in Unsloth Desktop with thinking controls:
Params
39.8 B
Context
262,144
Downloads 30d
308 K
Likes
372
Download history
daily snapshots · 14 days308 K150 K
Oct 28Nov 1Nov 6Nov 10
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 |
|---|---|---|---|---|
| flint-UD-IQ1_M.gguf | IQ1_M | 11.4 GB | 19.0 GB | ✅ Runs comfortably |
| flint-UD-IQ2_XXS.gguf | IQ2_XXS | 11.9 GB | 19.5 GB | ✅ Runs comfortably |
| flint-UD-Q2_K_XL.gguf | Q2_K_XL | 13.6 GB | 21.4 GB | ⚠️ Tight — reduce context |
| flint-UD-IQ3_XXS.gguf | IQ3_XXS | 14.5 GB | 22.5 GB | ⚠️ Tight — reduce context |
| flint-UD-Q3_K_XL.gguf | Q3_K_XL | 18.5 GB | 26.8 GB | ❌ Won’t fit |
| flint-UD-Q4_K_XL.gguf | Q4_K_XL | 25.1 GB | 34.0 GB | ❌ Won’t fit |
| flint-Q8_0.gguf | Q8_0 | 42.3 GB | 53.0 GB | ❌ Won’t fit |
| flint-BF16-00002-of-00003.gguf | GGUF | 49.9 GB | 61.4 GB | ❌ Won’t fit |
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$ ollama run flint-gguf # pin the quantization explicitly $ ollama run flint-gguf-q4_k_xl
$ huggingface-cli download unsloth/flint-GGUF-GGUF \
flint-UD-Q4_K_XL.gguf --local-dir .
$ llama-cli -m flint-UD-Q4_K_XL.gguf \
-c 8192 -ngl 99 -t 8 --color
$ curl -s https://aimodelscomparison.com/api/v1/models/flint-gguf
{
"hf_id": "unsloth/flint-GGUF",
"params_b": 39.79,
"context_length": 262144,
"license": { "id": "apache-2.0", "commercial": "yes" },
"downloads_30d": 308066,
"vram_estimates": [
{ "quant": "GGUF", "gb": 61.4 },
{ "quant": "Q8_0", "gb": 53.0 }
],
"updated_at": "2026-10-28T01:00:19Z"
}
Specifications
- Parameters
- 39.8 B
- Context length
- 262,144
- Licence
- apache-2.0
- First seen on the Hub
- 2026-10-20
- Base model
- flint
- Training datasets
- undisclosed
- Added to our catalog
- 2026-10-28
Compare with any text-generation model