Laguna-S-2.1-APEX-GGUF
MoE-aware, mixed-precision APEX quantization of poolside/Laguna-S-2.1 — a 118B-total / ~8B-active open-weight agentic coding MoE (48 layers, 256 routed + 1 shared expert, top-10 routing, GQA with per-head softplus output gating, mixed sliding-window/global attention, 1M context).
Params
117.6 B
Context
1,048,576
Downloads 30d
185 K
Likes
25
Download history
daily snapshots · 18 days190 K183 K
Oct 6Oct 12Oct 18Oct 23
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 |
|---|---|---|---|---|
| laguna-s-2.1-DFlash-Q4_K.gguf | Q4_K | 0.7 GB | 18.9 GB | ✅ Runs comfortably |
| laguna-s-2.1-DFlash-Q8_0.gguf | Q8_0 | 1.2 GB | 19.4 GB | ✅ Runs comfortably |
| Laguna-S-2.1-APEX-i-quality-v2.gguf | GGUF | 74.0 GB | 99.5 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 laguna-s-2-1-apex-gguf # pin the quantization explicitly $ ollama run laguna-s-2-1-apex-gguf-q4_k
$ huggingface-cli download Myric/Laguna-S-2.1-APEX-GGUF-GGUF \
laguna-s-2.1-DFlash-Q4_K.gguf --local-dir .
$ llama-cli -m laguna-s-2.1-DFlash-Q4_K.gguf \
-c 8192 -ngl 99 -t 8 --color
$ curl -s https://aimodelscomparison.com/api/v1/models/laguna-s-2-1-apex-gguf
{
"hf_id": "Myric/Laguna-S-2.1-APEX-GGUF",
"params_b": 117.56,
"context_length": 1048576,
"license": { "id": "openmdw-1.1", "commercial": "unknown" },
"downloads_30d": 184755,
"vram_estimates": [
{ "quant": "GGUF", "gb": 99.5 },
{ "quant": "Q4_K", "gb": 18.9 }
],
"updated_at": "2026-10-06T01:00:21Z"
}
Specifications
- Parameters
- 117.6 B
- Context length
- 1,048,576
- Licence
- openmdw-1.1
- First seen on the Hub
- 2026-07-22
- Base model
- Laguna-S-2.1
- Training datasets
- undisclosed
- Added to our catalog
- 2026-10-06
Family
Base model and the most-downloaded derivatives in the catalog.
Compare with any text-generation model