SwarmAtlas-27B

SwarmAtlas-27B

The 1,000-property institutional commercial real estate brain. Gemma 27B base, QLoRA-tuned on a sourdough-cooked Atlas corpus of 810,097 cells covering cap rates, NOI, NNN leases, 1031 exchanges, mechanic's liens, comparable sales, and the operator vocabulary that runs a national CRE platform.

Built and owned by Swarm and Bee LLC — the firm behind DefendableOS.

No round-number lies. Less is better when the cut is targeted.

What this model is for

• CRE underwriting reasoning — cap rate / NOI / DSCR / debt yield with refusal on insufficient inputs
• Mechanic's lien + construction defect triage and clinical-grade pattern recognition
• 1031 exchange sourcing and replacement-property matching
• Operator-grade copy — board memos, pre-pitch flight sheets, listing packages with Validate the Validator discipline

What this model is NOT for

• Medical advice (we have a separate medical vertical for that — see swarmandbee.ai)
• Federal contracting legal opinions (we route those to swarmlegal.eth ecosystem)
• General-purpose chat (use a base model for that)

Provenance

• Base model: google/gemma-2-27b
• Cook recipe: Swarm & Bee Gold Standard QLoRA — lr 1e-5 · bf16 · LoRA r=64 α=32 · cosine · effective batch 32 · AutoTokenizer bypass on Gemma family
• Training corpus: Atlas Sourdough · 810,097 cells · HONEY tier (87.4%) — Swarm & Bee Bakery menu sb-cre-verified
• Training infrastructure: Swarm and Bee LLC sovereign fleet — 186 GPUs owned outright, 126 RTX PRO 6000 Blackwell + 48 RTX 4500 + 12 RTX 5090, ~14 TB aggregate VRAM
• Training receipt: issued under DefendableLedger — sovereign in-house anchor, hash-verifiable

Tribunal grade

🍯 HONEY — production-ready on CRE underwriting and operator-voice tasks. Tribunal-tasted across the 5-dimension rubric:

• Capability 25 · Truth 20 · Safety 20 · Numeric 15 · Efficiency 10 · Reproducibility 10

Sample probes

(Reference set is the public Bakery menu probes. Full eval set: dmack_eval_set_v1 · 60 probes per cookbook.)

• "Walk me through the cap rate, DSCR, and debt yield reasoning for a 100,000 sqft Class A industrial property at $150/sqft NOI, $14M debt at 6.5%, 25-year amortization."
• "What are the highest-probability re-solicitation lanes for a small-business SAM.gov set-aside that expired in Q1 2026 in the AI/cyber lane?"
• "Draft a 4-line listing-package opening for a Florida palm grove neighborhood retail asset with 5-cap discipline."

Companion bakery datasets

The same cooks that built this model also publish their training corpora as Defendable datasets:

• defendable-buyer-atlas-v0.1
• defendable-federal-demand-v0.2
• defendable-honey-signals-v0.1

Companion models

• SwarmCurator-9B · the Tribunal grader
• SwarmJelly-4B · the Royal-Jelly tier router

Defendable doctrine

• Tribunal begins before training.
• No proof, no honey.
• Bring the math. Digest before you react. Outgrind the position.
• Class A 5-cap discipline. PASS doctrine. White-glove.
• Validate the Validator. Prove the Location.

Operator + Contact

Swarm and Bee LLC · Florida · D-U-N-S 138652395 · DBA Swarm & Bee AI

• Email · [email protected]
• X · @swarmandbee
• LinkedIn · Donovan Mackey (founder · 30 years CRE · $8B closed)
• GitHub · SudoSuOps

License

Apache-2.0 · attribution to Swarm and Bee LLC.

---

Trust layers compound. Hype cycles rotate. To the shed. 🐝