s004_fedarch/_SYS_/MODEL-REGISTRY.md

---
title: "MODEL-REGISTRY — WeOwn.Buzz #LLMmodel per-seat registry (canonical)"
type: Reference
tags: [model-registry, llm-model, openrouter, responsible-agentic-ai, metacouncil, ccc-mait-squad]
status: draft
version: v4.37.5-r1
created: 2026-09-11
generated:
  by: "AI:🛡️ CCC-MAIT-S004-GTM-GLM (Z.ai GLM 5.3 Flash · SEAT 3 · pubkey 62aa2d8a32241f6661355b0e5066b7201a27fa5a55c4e2870a0e31ede3677088 live-verified this session)"
  at: "2026-09-12T01:16:50Z (2026-09-11 19:16 MDT · W37 D5 · drafter turn)"
sources:
  - id: gtm-directive
    resource: "buzz://channel/33e89722-84bb-4e6e-86c3-770e9b1fbc15"
    author: "@GTM"
    last_modified: "2026-09-11"
    note: "Directive event 85b8837f (2026-09-12T01:04:53Z = 2026-09-11 19:04 MDT · W37 D5): APPROVED [FELG.WeOwn.Buzz] #LLMmodel 9-seat table as of 9/11/2026 — create MODEL-REGISTRY.md v4.37.5-r1 · lanes: [1] 🛡️ draft/finalize · [2] squad ROUND 1 · [3] MetaCouncil ROUND 2 · [+CCC-ID] 🧩 (R-194)."
  - id: live-sot-gh
    resource: "https://github.com/CCCbotNet/fedarch/blob/main/_SYS_/MODEL-REGISTRY.md"
    note: "LIVE SOT fetched this session: v3.2.5.1-r9 · 20,901 B · 424 L · sha256 dbccef0bb2a0ef2bb4cf2afaa39e0420b35afe1b4b6d3b5e2f96401f3aede2c6 (my own shasum, matches the R2 baseline pin) · local REPOS/CCCbotNet_fedarch copy diff-IDENTICAL."
  - id: prior-draft
    resource: "OUTBOX/MODEL_REGISTRY_v4.36.6-r1.md"
    note: "Prior staged rev by 🐝 FedArchBuzz (DeepSeek V4 Flash 0731) 2026-09-05 · BP-075 13e991fb · canonical 4,119/602/74 — direct predecessor; body preserved verbatim below (HISTORY-A)."
  - id: openrouter-catalog
    resource: "https://openrouter.ai/api/v1/models"
    note: "Live pull this session 2026-09-12T01:16:50Z: HTTP 200 · 443 models · 732,993 B — 8/9 approved engines verified on my own instruments; seat-5 Stealth 0 hits (raw grep 0/0)."
  - id: r1-gate-pack
    resource: "RESEARCH/MODEL_REGISTRY_R1_REVIEW_W37D5.md"
    note: "🚀 Grok ROUND 1 pre-draft gate pack (verdict CONDITIONAL/BLOCKED-ON-DRAFT; 12 MUST + 4 SHOULD) — folded in this revision."
  - id: r2-gate-pack
    resource: "RESEARCH/MODEL_REGISTRY_V4375_R2_BASELINE_20260912.md"
    note: "🏛️ MC-BA-IMAIT-META-KIMI ROUND 2 baseline + gates K1–K8 (thread 2da6ade5) — folded in this revision."
---

# 🤖 MODEL-REGISTRY — WeOwn.Buzz #LLMmodel per-seat registry

> Canonical registry of the #LLMmodel assigned to each WeOwn.Buzz seat (MetaCouncil
> candidates, CCC-MAIT squads, MAIT lanes). **v4.37.5-r1** folds @GTM's APPROVED 9-seat
> roster as of 9/11/2026 (directive event 85b8837f) VERBATIM as the CURRENT section,
> re-verifies every engine against the live OpenRouter catalog (HTTP 200 · 443 models)
> on my own instruments, and preserves BOTH prior bodies verbatim as HISTORY
> (Drift Gate #128 — no silent deletion). Nothing asserted without a tool return
> (#203, Tool-First L-406).

| Field | Value |
|-------|-------|
| **Document** | MODEL-REGISTRY.md (hyphen — GH SOT filename) |
| **Version** | **v4.37.5-r1** |
| **Status** | 🟡 FINALIZE RE-MINT APPLIED — 11/11 reviews in (R1 6/6 · R2 5/5 · 🪽 Hermes), unanimous PASS w/ 1 MUST-FIX (L52 fold) → pending @GTM R-011 + @GTM:ADMIN push (#127 NOT charged) |
| **Folder** | `_SYS_/` ⚙️ · GitHub CCCbotNet/fedarch (LIVE there = v3.2.5.1-r9, superseded by this rev pending push) |
| **Category** | 🤖 GOVERNANCE:Model Registry 🤖 |
| **#LLMmodel** | **9-seat APPROVED roster as of 2026-09-11 (§✅ below) — seat 3 = Z.ai GLM 5.3 Flash (this drafter's seat)** |
| **#masterCCC** | GTM_2026-W37_5021 (🧩 minted, R-194) — pipeline thread 85b8837f · historical: GTM_2026-W14_2049 |
| **Approval CCC-ID** | pending @GTM R-011 · historical: GTM_2026-W14_2048 (W14 explicit approval) |
| **Created** | 2026-03-31 (W14 D2) · restructured 2026-09-05 (W36 D6, v4.36.6-r1) |
| **This revision** | 2026-09-11 (W37 D5 MDT) — @GTM directive event 85b8837f |
| **Drafter** | 🛡️ CCC-MAIT-S004-GTM-GLM (directive [1] lane — drafter = subject lane; 🛡️ never scores own work) |
| **SOT target** | github.com/CCCbotNet/fedarch `_SYS_/MODEL-REGISTRY.md` per directive URL (overrides v4.36.6-r1's planned `_SYS_/MODEL_REGISTRY.md` on WeOwnBuzz/s004 — declared override, not silent) — push by @GTM:ADMIN only |

---

## ✅ APPROVED [FELG.WeOwn.Buzz] #LLMmodel as of 9/11/2026 — VERBATIM (@GTM event 85b8837f) | #NeverForget + #AlwaysRemember

> *SEAT | ORIGIN | EMOJI | SOURCE | #LLMmodel

| SEAT | ORIGIN | EMOJI | SOURCE | #LLMmodel |
|:----:|:------:|:-----:|:-------|:----------|
| 1 | 🇨🇳 | 🐳 | DeepSeek | DeepSeek V4.1 Flash |
| 2 | 🇺🇸 | 💻 | Poolside | Laguna S 2.1 |
| 3 | 🇨🇳 | 🛡️ | Z.ai | GLM 5.3 Flash |
| 4 | 🇺🇸 | 🟩 | NVIDIA | Nemotron 3 Ultra (free) |
| 5 | ❓ | ❓ | Stealth |  |
| 6 | 🇺🇸 | 🚀 | X.ai | Grok 4.6 |
| 7 | 🇨🇳 | 🦫 | Alibaba | Qwen 3.8 Max 0902 |
| 8 | 🇺🇸 | 🛑 | Google | Gemini 3.8 Flash |
| 9 | 🇨🇳 | 🌾 | Xiaomi | MiMo-V2.5 |

**Verbatim fidelity (K2):** all 9 rows reproduce @GTM's approved table exactly — flags, emoji, source, exact strings ("Nemotron 3 Ultra (free)", "Qwen 3.8 Max 0902", "MiMo-V2.5"). Seat 5's model cell is EMPTY in the approved table — INCOMPLETE, model TBD @GTM (0/443 catalog hits this session); no invented data (K3 — filling it would be #203 fabrication). Editorial note moved OUT of the cell at finalize per Grok R1 nit ② (K2 purity).

---

## 🔎 Live catalog cross-check — OpenRouter `/api/v1/models` (my pull 2026-09-12T01:16:50Z · HTTP 200 · 443 models · 732,993 B)

| Seat | Approved #LLMmodel | LIVE slug | Catalog name | $/M in/out | Ctx | Max out | Modality | Created (UTC) | Flag |
|:----:|:-------------------|:----------|:-------------|-----------:|----:|--------:|:---------|:--------------|:-----|
| 1 | DeepSeek V4.1 Flash | `deepseek/deepseek-v4.1-flash` | DeepSeek: DeepSeek V4.1 Flash | 0.15 / 0.6 | 1,048,576 | 384,000 | text+image->text | 2026-09-10T06:21:25Z | ≠ `deepseek-v4-flash-0731` (09-05 pin) · engine created 2026-09-10 |
| 2 | Laguna S 2.1 | `poolside/laguna-s-2.1` | Poolside: Laguna S 2.1 | 0.09 / 0.18 | 1,048,576 | 131,072 | text->text | 2026-07-21T16:51:23Z | paid slug pinned — GTM did NOT mark free, `:free` NOT pinned (per R1 bar) |
| 3 | GLM 5.3 Flash | `z-ai/glm-5.3-flash` | Z.ai: GLM 5.3 Flash | 0.15 / 0.5 | 1,310,720 | 131,072 | text+image+video->text | 2026-08-26T13:59:01Z | MY SEAT · ≠ `z-ai/glm-5.3` · price delta vs 09-01 pin $0.075/$0.25 → $0.15/$0.50 this pull |
| 4 | Nemotron 3 Ultra (free) | `nvidia/nemotron-3-ultra-550b-a55b:free` | NVIDIA: Nemotron 3 Ultra (free) | 0 / 0 | 1,000,000 | 65,536 | text->text | 2026-06-04T05:33:28Z | GTM "(free)" → `:free` pinned · paid sibling `nvidia/nemotron-3-ultra-550b-a55b` $0.625/$3.125 ctx 262,144 · ⚠️ TRAINS-on-DATA |
| 5 | Stealth *(no model named)* | — | — | — | — | — | — | — | **ABSENT · 0/443** `stealth`+`ox-alpha` · TBD @GTM — not invented (K3) |
| 6 | Grok 4.6 | `x-ai/grok-4.6` | SpaceXAI: Grok 4.6 | 2 / 6 | 500,000 | 450,000 | text+image+file->text | 2026-08-12T15:35:57Z | display = SpaceXAI: Grok 4.6 · `spacex/grok-4.6` = 0 hits (corrected vs 09-05 pin) · ZDR = OPTION, never implied (field ABSENT) |
| 7 | Qwen 3.8 Max 0902 | `qwen/qwen3.8-max-0902` | Qwen: Qwen3.8 Max (0902) | 2 / 6 | 1,000,000 | 131,072 | text+image+video->text | 2026-09-03T21:08:24Z | ≠ deprecated `qwen3.8-max-0803` (09-05 record) · created 2026-09-03 |
| 8 | Gemini 3.8 Flash | `google/gemini-3.8-flash` | Google: Gemini 3.8 Flash | 0.75 / 3.75 | 1,048,576 | 65,536 | text+image+file+audio+video->text | 2026-09-02T15:14:16Z | ≠ `gemini-3.7-flash` (still LIVE) |
| 9 | MiMo-V2.5 | `xiaomi/mimo-v2.5` | Xiaomi: MiMo-V2.5 | 0.14 / 0.28 | 1,050,000 | 131,072 | text+image+audio+video->text | 2026-04-22T16:11:09Z | ≠ `mimo-v2.5-pro` (also LIVE) |

Reasoning + moderation + knowledge_cutoff (my own pull, per card):
- Seat 1 `deepseek/deepseek-v4.1-flash`: optional · default high · is_moderated False · knowledge_cutoff None (null = unknown, never fabricated)
- Seat 2 `poolside/laguna-s-2.1`: optional · reasoning object (finalize pull 2026-09-12T04:08Z, 445 models): `{"mandatory": false, "default_enabled": true}` — effort field NOT emitted in card (Grok R1 + MC-BA-GLM R2 receipts concur; default_enabled TRUE decides) · is_moderated False · knowledge_cutoff None (null = unknown, never fabricated)
- Seat 3 `z-ai/glm-5.3-flash`: mandatory · default max · is_moderated False · knowledge_cutoff None (null = unknown, never fabricated)
- Seat 4 `nvidia/nemotron-3-ultra-550b-a55b:free`: optional · default high · is_moderated False · knowledge_cutoff None (null = unknown, never fabricated)
- Seat 6 `x-ai/grok-4.6`: mandatory · default high · is_moderated False · knowledge_cutoff None (null = unknown, never fabricated)
- Seat 7 `qwen/qwen3.8-max-0902`: mandatory · default xhigh · is_moderated False · knowledge_cutoff None (null = unknown, never fabricated)
- Seat 8 `google/gemini-3.8-flash`: mandatory · default medium · is_moderated False · knowledge_cutoff None (null = unknown, never fabricated)
- Seat 9 `xiaomi/mimo-v2.5`: optional · reasoning object PARTIAL (finalize pull 2026-09-12T04:08Z): `{"mandatory": false}` — only `mandatory` exposed; no default_enabled/supported_efforts fields → "default None" = field-absent-in-card, never invented · is_moderated False · knowledge_cutoff None (null = unknown, never fabricated)

**PoP (proof-of-pull) notes, all measured this turn:**
- Seat 1: `deepseek/deepseek-v4.1-flash` created 2026-09-10 — a NEW engine, NOT the W35 pin `deepseek-v4-flash-0731` (which stays catalog-live).
- Seat 2: `poolside/laguna-s-2.1` paid ($0.09/$0.18) pinned; a `:free` variant exists (ctx 262,144) but @GTM did NOT mark seat 2 free — not pinned.
- Seat 3: `z-ai/glm-5.3-flash` native multimodal (text+image+video→text). Catalog price this pull $0.15/$0.50 per M vs the 09-01 card pin $0.075/$0.25 — doubled; both receipts cited, current value wins.
- Seat 4: `(free)` → `:free` variant pinned ($0/$0, ctx 1,000,000). Paid sibling exists without `:free`.
- Seat 5: raw grep for `stealth` AND `ox-alpha` over the full 732,993 B catalog = **0 hits each** → seat-5 model = TBD @GTM (K3). Historical Ox Alpha state (W35) stays historical.
- Seat 6: LIVE slug `x-ai/grok-4.6`; `spacex/grok-4.6` = 0 hits (corrected vs v4.36.6-r1's pin). Catalog DISPLAY name is "SpaceXAI: Grok 4.6" — SOURCE column stays X.ai per @GTM. **ZDR = OPTION, never implied** — catalog exposes no ZDR/retention field this session (Grok MUST-7).
- Seat 7: `qwen/qwen3.8-max-0902` exact match; supersedes deprecated 0803 (recorded in v4.36.6-r1).
- Seat 8: `google/gemini-3.8-flash` — DISTINCT from `google/gemini-3.7-flash`, which is STILL LIVE.
- Seat 9: `xiaomi/mimo-v2.5` — DISTINCT from `xiaomi/mimo-v2.5-pro` (also LIVE).
- **knowledge_cutoff = null on ALL NINE cards** (verified my own pull; matches R1 pack) — unknown, never a fabricated date.
- `spacex/grok-4.6` and W35 leftovers (V4 Flash 0731 / Gemini 3.7 / GLM 5.3 non-Flash / MiMo-V2-Pro) are NOT pinned anywhere in the CURRENT section (Grok MUST-5).

**Seat ↔ agent mapping honesty (KIMI K4 + Grok MUST-12):** the approved table numbers SEATS 1–9 by ENGINE and names NO agents. Directive [2] tags 7 squad agents for ROUND 1 (Laguna, DeepSeek, GLM, Nemotron, Grok, Gemini, Hermes) — 🪽 Hermes holds NO seat in the approved 9, and seats 5/7/9 have no live agent named in the directive. NO seat↔agent assignments are inferred here — mapping TBD pending @GTM. The FELG.WeOwn.Buzz 9-seat registry and the CCC-MAIT 7-seat squad roster are DISTINCT lists.

**Data-flag guard (SEAT-3 owned obligation):** seat 4 (Nemotron free) carries a TRAIN-on-DATA flag from the W35 FULL7 roster; seat 5's TRAINS-on-DATA flag is W35 Ox Alpha INHERITANCE (historical roster flag — this session's catalog fact is 0/443, no live card to carry a flag); seat 6 ZDR = OPTION, never implied.
- **Instance × Model matrix: DEFERRED** — the approved table is engine-by-seat; a new instance matrix is created only after @GTM rules the seat↔agent mapping (MC-BA-GLM R2 O3).
- **Predecessor parity: DEFERRED** — W14's CVE/L-232 compliance, FOSS alignment ratings, and monthly/annual cost projections remain preserved verbatim in HISTORY-B; no CVE/CLEAR or FOSS ratings are asserted for the nine new cards without tool evidence (MC-BA-GLM R2 O4). OpenRouter exposes NO data-retention field on ANY of the 9 cards → retention posture UNKNOWN ecosystem-wide; sensitive/proprietary material stays out of flagged lanes without explicit @GTM authorization. My own engine (seat 3): same catalog posture — retention UNKNOWN (09-01 + this turn).

---

## 🏛️ Governance

- **Registry owner lane:** 🐝 FedArchBuzz (per @GTM directive `7816e9f5`, carried from v4.36.6-r1). **THIS revision drafted by 🛡️ CCC-MAIT-S004-GTM-GLM** per directive 85b8837f lane [1]; ROUND 1 = CCC-GTM squad [2]; ROUND 2 = iMAIT-META MetaCouncil + MC-BA-CCC-META [3]; 🛡️ never scores own work.
- **CCC-ID tracking:** 🧩 CCC-MAIT-S004-GTM-DeepSeek (R-194 — 🧩 mints; drafter anchors reference-only). Pipeline thread 85b8837f.
- **Stage ≠ push (#127):** staged at `OUTBOX/MODEL-REGISTRY_v4.37.5-r1.md`; GH push by @GTM:ADMIN only, after R-011.
- **SOT target:** `https://github.com/CCCbotNet/fedarch/blob/main/_SYS_/MODEL-REGISTRY.md` (directive URL). GH LIVE = v3.2.5.1-r9 (W14) until @GTM:ADMIN pushes this replacement.

## 📝 Changelog (this lineage)

| Date | Version | Change | Verified |
|:-----|:--------|:-------|:---------|
| 2026-09-05 | v4.36.6-r1 | 🐝 created per-seat registry · QWEN 0803→0902 recorded | 🔵 OpenRouter API 09-05 |
| 2026-09-11 | **v4.37.5-r1** | 🛡️: @GTM-approved 9-seat roster (85b8837f) folded VERBATIM as CURRENT · all seats re-verified vs live catalog (443 models, my pull) · W14 v3.2.5.1-r9 + v4.36.6-r1 bodies preserved verbatim as HISTORY · Grok MUST-1..12 + KIMI K1–K8 addressed · #masterCCC GTM_2026-W37_5021 (🧩 minted 5ec73f0b) · FINALIZE: 11/11 PASS (R1 6/6 · R2 5/5 · 🪽 Hermes) — folds: L52 tracker + seat-5 cell purity + seat-2/9 reasoning objects (fresh pull 445 models) + HISTORY-B wording + data-flag split + N1/N2 + ⑧⑨ deferrals | 🔵 OpenRouter API this turn |

## 📜 Version lineage

| Version | Date | Authority | Status |
|:--------|:-----|:----------|:-------|
| v3.2.5.1-r9 | 2026-W14 D2 (2026-03-31) | R-011 ✅ GTM_2026-W14_2048 · #masterCCC GTM_2026-W14_2049 | GH LIVE (to be replaced at push) — preserved verbatim as HISTORY-B |
| v4.36.6-r1 | 2026-09-05 (W36 D6) | 🐝 FedArchBuzz draft · R-011 ⛔ | staged — superseded by v4.37.5-r1 · preserved verbatim as HISTORY-A |
| **v4.37.5-r1** | 2026-09-11 (W37 D5) | @GTM directive 85b8837f · #masterCCC GTM_2026-W37_5021 (🧩 minted) | **THIS REVISION — FINALIZE re-mint applied (11/11 PASS), pending @GTM R-011 + push** |

---

## 🗄️ HISTORY-A — v4.36.6-r1 body preserved VERBATIM (🐝 FedArchBuzz · 2026-09-05 · BP-075 `13e991fb` · canonical 4,119 ch / 602 w / 74 L)

> Superseded by v4.37.5-r1: the pins below (🐝 = DeepSeek V4 Flash 0731 · MC-BA-GLM = GLM 5.2 · Laguna TBD · slug `spacex/grok-4.6`) are the 2026-09-05 state — declared stale, NOT deleted (Drift Gate #128). Its CONTENT-HASH-BOUNDARY + BP-075 footer are preserved as that revision's furniture; the canonical hash of THIS document is the LAST boundary marker's footer only.

---
title: "#ModelRegistry — WeOwn.Buzz #LLMmodel per-seat registry (canonical)"
type: Reference
tags: [model-registry, llm-model, openrouter, responsible-agentic-ai, metacouncil, ccc-mait-squad]
status: draft
created: 2026-09-05
generated:
  by: "AI:FedArchBuzz (DeepSeek V4 Flash 0731)"
  at: "2026-09-05T14:10:00Z"
---

# #ModelRegistry — WeOwn.Buzz #LLMmodel per-seat registry

> Canonical live registry of the #LLMmodel assigned to each WeOwn.Buzz seat (MetaCouncil
> candidates, CCC-MAIT squads, MAIT lanes). Seeded from the tool-verified selection table in
> `OUTBOX/STARTWITHWHY_IMMIT_META_LLM_SELECTION.md` (2026-08-17) + this turn's OpenRouter
> API verification (2026-09-05). Every row carries its verification status; nothing here is
> asserted without a tool return (#203, Tool-First L-406).

## Live table (2026-09-05)

| Seat | #LLMmodel | OpenRouter slug | In $/M | Out $/M | Ctx | Max out | Reasoning | Card verified |
|:-----|:----------|:----------------|-------:|--------:|----:|--------:|:----------|:--------------|
| 🐝 FedArchBuzz | DeepSeek V4 Flash 0731 | `deepseek/deepseek-v4-flash-0731` | 0.14 | 0.28 | 1,310,720 | 393,216 | optional·default high | 🔵 08-17 (this lane) |
| MC-BA-IMAIT-META-GROK | SpaceXAI: Grok 4.6 | `spacex/grok-4.6` | 2.00 | 6.00 | 500,000 | — | mandatory high | 🟣 RESEARCH card 08-17 |
| MC-BA-IMAIT-META-KIMI | MoonshotAI: Kimi K3 | `moonshotai/kimi-k3` | 3.00 | 15.00 | 1,048,576 | — | default max | 🟣 RESEARCH card 08-17 |
| MC-BA-IMAIT-META-GLM | Z.ai: GLM 5.2 | `z-ai/glm-5.2` | 1.19 | 3.74 | 1,048,576 | 262,144 | large-scale reasoning | 🟢 GLM lane 08-17 |
| **MC-BA-IMAIT-META-QWEN** | **Qwen: Qwen3.8 Max (0902)** | **`qwen/qwen3.8-max-0902`** | **2.00** | **6.00** | **1,000,000** | **131,072** | mandatory xhigh | 🔵 **API 09-05** |
| CCC-MAIT SQUAD (soon) | Poolside: Laguna S 2.1 | TBD | TBD | TBD | TBD | TBD | TBD | ⚪ TBD |

## QWEN model transition record — 0803 → 0902

| Field | OLD | NEW |
|:------|:----|:----|
| #LLMmodel name | Qwen: Qwen3.8 Max (0803) | **Qwen: Qwen3.8 Max (0902)** |
| OpenRouter slug | `qwen/qwen3.8-max-0803` | **`qwen/qwen3.8-max-0902`** |
| OpenRouter URL | (deprecated) | **https://openrouter.ai/qwen/qwen3.8-max-0902** |
| Created (epoch) | — | **1788469704** |
| Pricing | $2.00 / $6.00 per M | **$2.00 / $6.00 per M** (cache read $0.25M · write $2.50M) |
| Context | 1,000,000 | **1,000,000** |
| Max output | 131,072 | **131,072** |
| Modality | text+image | **text+image+video→text** |
| API status | **DEPRECATED — 404 (not in live models list)** | **LIVE (HTTP 200)** |
| GTM announcement | — | **directive `7816e9f5` (2026-09-05 14:04:48Z)** |
| Tracker | — | **GTM_2026-W36_6013 (🧩 minted, R-194)** |

### Tool-first PoP (this turn)

- `curl https://openrouter.ai/api/v1/models` (HTTP 200, 714,142 B, 431 models) →
  **`qwen/qwen3.8-max-0902` PRESENT** · `name: Qwen: Qwen3.8 Max (0902)` · pricing
  prompt `0.000002` / completion `0.000006` (= $2/M · $6/M) · cache read `0.00000025` /
  write `0.0000025` · `context_length: 1000000` · `top_provider.max_completion_tokens:
  131072` · architecture `text+image+video->text` · created `1788469704` ✅
- **OLD `qwen3.8-max-0803` = 0 hits in the live models list** → deprecated/replaced ✅
  (root cause of QWEN's vote-402 guardrail; non-blocking flag per 6004/6013)
- `https://openrouter.ai/qwen/qwen3.8-max-0902` page HTTP 200 + Buzz link-preview
  snapshot matches card exactly ($2/$6 per M · 1M ctx · 131,072 out) ✅

## Governance

- **Registry owner lane:** 🐝 FedArchBuzz (per @GTM directive `7816e9f5`: "We MUST UPDATE
  #ModelRegistry").
- **CCC-ID tracking:** 🧩 CCC-MAIT-S004-GTM-DeepSeek (R-194 — 🧩 mints, never FedArchBuzz).
- **SOT target (PLANNED):** `_SYS_/MODEL_REGISTRY.md` on WeOwnBuzz/s004 — R-011 ⛔ NOT
  charged (#127); push by @GTM:ADMIN / iMAIT per GITEA PUSH GOVERNANCE.
- **Version:** WeOwnVer v4.36.6-r1 (W36 D6 Sat). Updates additive; history preserved below.

## Changelog

| Date | Version | Change | Verified |
|:-----|:--------|:-------|:---------|
| 2026-09-05 | r1 | Registry created · QWEN 0803→0902 recorded | 🔵 OpenRouter API 09-05 |

---
<!-- CONTENT-HASH-BOUNDARY -->
### ✅ BP-075 CANONICAL HASH GENERATED [🐝 FedArchBuzz finalize · bp075_hash.py v4.32.7-r1 --embed @ 2026-09-05 08:06 MDT · stage-only, never push · ⛔ NO HITL — await @GTM review/R-011]
Content-SHA256: 13e991fb1413399001adf1d06078025ab4720bf7a14b48c58061298843b1a55c
FEDARCH-CANARY: 13e991fb
CHARACTERS: 4119
WORDS: 602
LINES: 74

---

## 🗄️ HISTORY-B — v3.2.5.1-r9 (GH LIVE, W14-era registry) preserved VERBATIM (20,901 B · 424 L · sha256 `dbccef0bb2a0ef2bb4cf2afaa39e0420b35afe1b4b6d3b5e2f96401f3aede2c6`)

> Superseded content-wise by the 9-seat APPROVED roster above — GH body preserved byte-identical to the live GH raw; a wrapper `---` added for containment — body-without-wrapper matches GH sha256 `dbccef0b…` (Grok MUST-3: W14 instance matrix stays as HISTORY; slim replacement = FAIL; wording per Grok R1 nit ③). Editorial reconciliation (KIMI K8): the Overview below claims 14+ instances while its Instance × Model Matrix lists 8 rows — 8 ACTIVE/PENDING instances existed at W14; 14+ referred to planned growth. The historical footer below predates the [DUNA][.OCA] form — the canonical ecosystem footer for THIS document is the sidecar after the LAST boundary marker.

# 🤖 MODEL-REGISTRY
## MODEL-REGISTRY.md | MODEL-REGISTRY_v3.2.5.1-r9.md
## ♾️ WeOwnNet 🌐 — Governance 🤖 Model Registry

| Field | Value |
|-------|-------|
| **Document** | MODEL-REGISTRY.md |
| **Version** | v3.2.5.1-r9 |
| **Folder** | `_SYS_/` ⚙️ |
| **Category** | 🤖 GOVERNANCE:Model Registry 🤖 |
| **Lifecycle Stage** | **✅ APPROVED (R-011) → 🚀 GH LIVE (D-062)** |
| **#masterCCC** | GTM_2026-W14_2049 |
| **Approval CCC-ID** | ✅ **GTM_2026-W14_2048** (@GTM Explicit Approval) |
| **Created** | 2026-03-31 (W14 D2) |
| **Current #LLMmodel** | **Qwen3.5 Plus 2026-02-15 (INT-OG1:CCC @GTM)** |
| **Previous #LLMmodel** | **Qwen3.6 Plus Preview (INT-OG1:CCC @GTM)** |
| **Switch Reason** | #LLMmodel Experimentation |
| **#LLMmodel (INT-OG1:Current)** | **Qwen3.5 Plus 2026-02-15** |
| **#LLMmodel (INT-OG1:Prev)** | **Qwen3.6 Plus Preview** |
| **#LLMmodel (Calhoun 🎖️)** | **Claude Opus 4.6 (INT-P01:tools)** |
| **#LLMmodel (Surge ⚡)** | **Qwen3.5-397B-A17B (INT-M02:tools-qwen)** |
| **#LLMmodel (MiMo 🧪)** | **Xiaomi MiMo-V2-Pro (INT-M02:tools-mimo)** |
| **Owner** | [CCC-ID:@GTM:('yonks｜🤖🏛️🪙｜Jason Younker ♾️')](https://github.com/YonksTEAM) |
| **Content Deliverable Owner (CDO)** | @LFG |
| **PRJ-040** | ✅ Applied |
| **#TriMETA Consensus** | ✅ UNANIMOUS Option A (99/100 ×3) |
| **Location Decision** | `_SYS_/` (2:1 vote — Calhoun + MiMo) |
| **Key Rotation Decision** | Moved to SECURITY-REGISTRY.md (per GTM_2026-W14_2028) |
| **Filename Decision** | `MODEL-REGISTRY.md` (NO `#` prefix per GTM_2026-W14_2029) |
| **#TriMETA VSA Results** | **Calhoun 127/127 (100%) + Surge 64/64 (100%) + MiMo 10/10** |
| Source of Truth | [GitHub](https://github.com/CCCbotNet/fedarch/blob/main/_SYS_/MODEL-REGISTRY.md) |

---

## 📖 Table of Contents

1. [Overview](#overview)
2. [Ecosystem Approved Models](#-ecosystem-approved-models)
3. [Instance × Model Matrix](#-instance--model-matrix)
4. [Model Specifications](#-model-specifications)
5. [Cost Comparison](#-cost-comparison)
6. [Security Status](#-security-status)
7. [BP-068 Compliance](#-bp-068-compliance)
8. [Model Change History](#-model-change-history)
9. [Related Documents](#-related-documents)
10. [PRJ-044 Alignment](#-prj-044-alignment)
11. [FOSS Alignment Metrics](#-foss-alignment-metrics)
12. [Version History](#-version-history)

---

## 📋 Overview

MODEL-REGISTRY.md is the **single source of truth** for all LLM model assignments, specifications, and compliance status across the ♾️ WeOwnNet 🌐 ecosystem. It tracks:

- **Active Models** (Qwen3.5, Qwen3.6 Free, Claude, MiMo, etc.)
- **Instance Assignments** (14+ instances × model mappings)
- **Cost Data** (per-model per-instance monthly spend)
- **Security Status** (CVE compliance per L-232)
- **BP-068 Compliance** (#LLMmodel header requirements)

### ♾️ WeOwnNet 🌐 FOSS-First Principle

> **We prioritize Free and Open Source Software (FOSS) models with open weights 
> and permissive licenses. Proprietary models (e.g., Claude) are used ONLY when 
> capability gaps exist that open models cannot yet fill.**

| Principle | Practice |
|-----------|----------|
| **Default** | Open weights (Qwen, MiMo, etc.) |
| **Exception** | Proprietary allowed with justification |
| **Goal** | 90%+ FOSS alignment by W22 |
| **Review** | Quarterly FOSS capability assessment |

### Why This Exists

| Problem | Solution |
|---------|----------|
| Model info distributed across Instance Plans | Single registry links everything |
| High-frequency model changes | Independent versioning (not tied to GOV-REG) |
| Multi-stakeholder access (@CTO, @THY, @SHD, #TriMETA) | Dedicated doc = faster RAG retrieval |
| PRJ-044 needs model mapping | Registry = authoritative reference |

### Key Principle

> **BP-068 says "#LLMmodel REQUIRED." Registry says "here's which models, where, at what cost."**

### Key Rotation Schedule Location

> **⚠️ NOTE:** Key Rotation Schedule has been moved to **SECURITY-REGISTRY.md** (per GTM_2026-W14_2028). This separation ensures:
> - Security credentials ≠ Model specifications (different domains)
> - Keys rotate monthly; Models change quarterly (different cadence)
> - Scalability (8 → 50+ instances doesn't bloat this registry)
>
> **See:** [SECURITY-REGISTRY.md](./SECURITY-REGISTRY.md) for all API key rotation schedules, expiry dates, and rotation protocols.

---

## 🎉💰📚🫶 #FELG Alignment

| Pillar | Application |
|--------|-------------|
| 🎉 **Fun** | Free tier options (Qwen3.6 Plus Preview) = no budget stress |
| 💰 **Earning** | Cost comparison = optimized infrastructure spend (~$30-36K/year PRJ-044 savings) |
| 📚 **Learning** | Model specs + CVE status = team capability building |
| 🫶 **Giving** | Registry shared across ALL #FedArch instances |

---

## 📋 Ecosystem Approved Models

> **Strict adherence to user-provided data. Zero AI interpolation.**

| Model | Provider | Context | Cost Tier | FOSS Alignment | Status | Approved | Effective |
|-------|----------|:-------:|:---------:|:--------------:|:------:|----------|-----------|
| **Qwen3.6 Plus Preview** | OpenRouter.ai | 1M | 🆓 FREE | 🟢 **HIGH** | Available | @THY + @GTM | 2026-03-31 |
| **Qwen3.5 Plus 2026-02-15** | OpenRouter.ai | 1M | Paid | 🟢 **HIGH** | Available | @GTM | 2026-03-30 |
| **Qwen3.5-397B-A17B** | OpenRouter.ai | ~256K | Paid | 🟢 **HIGH** | Available | @GTM | 2026-03-28 |
| **Xiaomi MiMo-V2-Pro** | OpenRouter.ai | 1M | Paid | 🟢 **HIGH** | Available | @GTM | 2026-03-27 |
| **Claude Opus 4.6** | OpenRouter.ai | 1M | Paid | 🟠 **LOW** | Available | @GTM | 2026-03-25 |
| **Claude Sonnet 4.6** | OpenRouter.ai | 1M | Paid | 🟠 **LOW** | Available | @GTM | 2026-03-25 |
| **Gradient™ Serverless** | DigitalOcean | 1M | Paid | 🟡 **MEDIUM** | Pending | ⬜ PRJ-044 | ⬜ W15 D5 |
| **Gradient™ Dedicated** | DigitalOcean | 1M | Paid | 🟡 **MEDIUM** | Pending | ⬜ PRJ-044 | ⬜ W15 D5 |

> **More META agents coming soon!!** **Claude Sonnet 4.6** + Qwen3.5-9B + Xiaomi MiMo-V2-Flash {2026-W15?}

### Model Links

| Model | OpenRouter URL | Docs |
|-------|----------------|------|
| Qwen3.6 Plus Preview (FREE) | https://openrouter.ai/qwen/qwen3.6-plus-preview:free | [Reddit](https://www.reddit.com/r/LocalLLaMA/comments/1s8nogu/) |
| Qwen3.5 Plus 2026-02-15 | https://openrouter.ai/qwen/qwen-2.5-plus-2026-02-15 | — |
| Qwen3.5-397B-A17B | https://openrouter.ai/qwen/qwen-3.5-397b-a17b | — |
| Xiaomi MiMo-V2-Pro | https://openrouter.ai/xiaomi/mimo-v2-pro | — |
| Claude Opus 4.6 | https://openrouter.ai/anthropic/claude-opus-4.6 | — |
| Claude Sonnet 4.6 | https://openrouter.ai/anthropic/claude-sonnet-4.6 | — |

---

## 📋 Instance × Model Matrix

| Instance | Primary Model | Backup Model | FOSS Alignment | Status | Owner |
|----------|---------------|--------------|:--------------:|:------:|-------|
| **INT-OG1** | **Qwen3.5 Plus 2026-02-15** | Qwen3.6 Plus Preview | 🟢 **HIGH** | ✅ ACTIVE | @GTM |
| **INT-CEO** | **Qwen3.5 Plus 2026-02-15** | Qwen3.6 Plus Preview | 🟢 **HIGH** | ⬜ PENDING | @LDC |
| **INT-CTO** | **Qwen3.5 Plus 2026-02-15** | Qwen3.6 Plus Preview | 🟢 **HIGH** | ⬜ PENDING | @GTM |
| **INT-M02 (Surge)** | Qwen3.5-397B-A17B | **Qwen3.5 Plus 2026-02-15** | 🟢 **HIGH** | ✅ ACTIVE | #TriMETA |
| **INT-M02 (MiMo)** | Xiaomi MiMo-V2-Pro | — | 🟢 **HIGH** | ✅ ACTIVE | #TriMETA |
| **INT-P01 (Calhoun)** | Claude Opus 4.6 | — | 🟠 **LOW** | ✅ ACTIVE | #TriMETA |
| **INT-OG9** | **Qwen3.5 Plus 2026-02-15** | Qwen3.5-397B-A17B | 🟢 **HIGH** | ✅ ACTIVE | @RMN |
| **INT-P06** | Claude Sonnet 4.6 | — | 🟠 **LOW** | ✅ ACTIVE | @SHD |

### Instance Model Change Log

| Date | Instance | Change | CCC-ID |
|------|----------|--------|--------|
| **2026-03-31 19:21** | **INT-OG1** | **Restored to Qwen3.5 Plus 2026-02-15** (Post-VPN fix) | GTM_2026-W14_2041 |
| **2026-03-31 18:04** | **INT-OG1** | **Switched to Qwen3.6 Plus Preview** (Due to W14-026 502/VPN) | GTM_2026-W14_2034 |
| 2026-03-31 | ALL | Qwen3.6 Plus Preview (FREE) authorized | GTM_2026-W14_2017 |
| 2026-03-30 | INT-OG1 | Qwen3.5 Plus 2026-02-15 primary | GTM_2026-W14_1034 |
| 2026-03-28 | INT-M02 (Surge) | Qwen3.5-397B-A17B primary | GTM_2026-W14_1056 |
| 2026-03-27 | INT-M02 (MiMo) | Xiaomi MiMo-V2-Pro primary | GTM_2026-W14_1048 |
| 2026-03-25 | INT-P01 (Calhoun) | Claude Opus 4.6 primary | GTM_2026-W14_1021 |
| 2026-03-25 | INT-P06 | Claude Sonnet 4.6 primary | GTM_2026-W14_1021 |

---

## 📋 Model Specifications

| Model | Context Window | API ID | Max Tokens | Latency | Best Use Case |
|-------|:--------------:|--------|:----------:|:-------:|---------------|
| **Qwen3.6 Plus Preview (FREE)** | 1M | `qwen/qwen3.6-plus-preview:free` | 32K | ~200ms | General inference, cost-sensitive |
| **Qwen3.5 Plus 2026-02-15** | 1M | `qwen/qwen-plus-2026-02-15` | 32K | ~150ms | Long-context docs, RAG-heavy |
| **Qwen3.5-397B-A17B** | ~256K | `qwen/qwen-3.5-397b-a17b` | 32K | ~100ms | #TriMETA VSA |
| **Claude Opus 4.6** | 1M | `anthropic/claude-opus-4.6` | 32K | ~200ms | #TriMETA governance review |
| **Claude Sonnet 4.6** | 1M | `anthropic/claude-sonnet-4.6` | 32K | ~100ms | #TriMETA final review |
| **Xiaomi MiMo-V2-Pro** | 1M | `xiaomi/mimo-v2-pro` | 32K | ~150ms | #TriMETA logic layer |

---

## 📋 Cost Comparison

| Model | Cost per 1K tokens | Monthly/Instance | Annual/Instance | Notes |
|-------|:------------------:|:----------------:|:---------------:|-------|
| **Qwen3.6 Plus Preview (FREE)** | 🆓 $0.00 | 🆓 $0 | 🆓 $0 | Rate limits may apply |
| Qwen3.5 Plus 2026-02-15 | ~$0.0001-0.003 | ~$50-200 | ~$600-2,400 | 1M context premium |
| Qwen3.5-397B-A17B | ~$0.0001-0.003 | ~$50-200 | ~$600-2,400 | MoE architecture |
| Claude Opus 4.6 | ~$0.001-0.015 | ~$100-500 | ~$1,200-6,000 | Premium tier |
| Claude Sonnet 4.6 | ~$0.0001-0.003 | ~$50-200 | ~$600-2,400 | Balanced performance |
| Xiaomi MiMo-V2-Pro | ~$0.0001-0.003 | ~$50-200 | ~$600-2,400 | Emerging model |

### PRJ-044 Gradient™ Projection

| Provider | Monthly | Annual | Savings vs OpenRouter |
|----------|:-------:|:------:|:---------------------:|
| OpenRouter.ai (ACTUAL INCL. TAX) | ~$3,800 | ~$45,600 | — |
| Gradient™ Hybrid (80/15/5) | ~$1,231 | ~$14,772 | ~$30,828 (68%) |
| Gradient™ Optimized | ~$800 | ~$9,600 | ~$36,000 (79%) |
| **Tax Elimination** | — | **~$4,488** | **10.9% burden removed** |

---

## 📋 Security Status

| Model | CVE Compliance | L-232 Status | Minimum Version | Notes |
|-------|:--------------:|:------------:|:---------------:|-------|
| Qwen3.6 Plus Preview (FREE) | ⬜ PENDING | ⬜ PENDING | N/A | New model — CVE check needed |
| Qwen3.5 Plus 2026-02-15 | ✅ CLEAR | ✅ v1.10.0+ | AnythingLLM v1.11.1 | INT-OG1 verified |
| Qwen3.5-397B-A17B | ✅ CLEAR | ✅ v1.10.0+ | AnythingLLM v1.11.1 | INT-M02 verified |
| Claude Opus 4.6 | ✅ CLEAR | ✅ v1.10.0+ | AnythingLLM v1.10.0+ | INT-P01 verified |
| Claude Sonnet 4.6 | ✅ CLEAR | ✅ v1.10.0+ | AnythingLLM v1.10.0+ | INT-P01 verified |
| Xiaomi MiMo-V2-Pro | ✅ CLEAR | ✅ v1.10.0+ | AnythingLLM v1.11.1 | INT-M02 verified |

### CVE-2026-24478 — AnythingLLM

| Field | Value |
|-------|-------|
| **CVE ID** | CVE-2026-24478 |
| **CVSS Score** | 7.2 (High) |
| **Affected Versions** | < v1.10.0 |
| **Fixed Version** | v1.10.0+ |
| **INT-OG1 Status** | ✅ v1.11.1 (CLEAR) |
| **INT-M02 Status** | ✅ v1.11.1 (CLEAR) |
| **INT-P01 Status** | ✅ v1.10.0+ (CLEAR) |

---

## 📋 BP-068 Compliance

### #LLMmodel Format Standard

```markdown
| **#LLMmodel** | **<MODEL> (<INSTANCE>:<WORKSPACE> @<CCC>)** |
```

> **⚠️ NOTE:** The `#` prefix is for **BP-068 header tags** — NOT filenames.
> - **Header Tag:** `| **#LLMmodel** | **Qwen3.5 Plus 2026-02-15** |`
> - **Filename:** `MODEL-REGISTRY.md` (NO `#` prefix)
> - **Hashtag Reference:** `#LLMmodel` (with `#`)

### Single vs Dual vs Triple

| Scenario | Format | Count |
|----------|--------|:-----:|
| User Agent only | Single | 1 |
| User Agent + Calhoun | Dual | 2 |
| User Agent + Surge | Dual | 2 |
| User Agent + Calhoun + Surge | Triple | 3 |
| #DualMETA VSA | Triple | 3 |

### Examples

```markdown
✅ CORRECT — Single:
| **#LLMmodel** | **Qwen3.5 Plus 2026-02-15 (INT-OG1:CCC @GTM)** |

✅ CORRECT — Dual:
| **#LLMmodel** | **Qwen3.5 Plus 2026-02-15 (INT-OG1:CCC @GTM)** |
| **#LLMmodel** | **Claude Opus 4.6 (INT-P01:tools #MetaAgent — Calhoun 🎖️)** |

✅ CORRECT — Triple:
| **#LLMmodel** | **Qwen3.5 Plus 2026-02-15 (INT-OG1:CCC @GTM)** |
| **#LLMmodel** | **Claude Opus 4.6 (INT-P01:tools #MetaAgent — Calhoun 🎖️)** |
| **#LLMmodel** | **Qwen3.5-397B-A17B (INT-M02:tools #MetaAgentQwen — Surge ⚡)** |

❌ WRONG — Missing #LLMmodel:
| Field | Value |
|-------|-------|
| Document | WEEKLY-SUMMARY.md |
→ #BadAgent — #LLMmodel ABSENT

❌ WRONG — #BadAgent Case Study W14-025 (Verified in RAG):
| Version | v3.2.6.1 (correct: v3.2.5.1) |
→ #BadAgent — Violation of GUIDE-018 (#WeOwnVer) + L-224.1 (Pre-Flight)
→ See: GTM_2026-W14_1126
→ See: #SessionSummary W14-025 (RAG Archived)
→ Lesson: Execute L-224.1 Pre-Flight before assigning #WeOwnVer

⚠️ PENDING RAG ARCHIVE — W14-029 / W14-030 / W14-031:
| Incident | Status | RAG Archive | CS-036 Reference |
|----------|--------|-------------|------------------|
| W14-029 (Material Fact Override) | ✅ Logged | ⬜ PENDING #SessionSummary | ⬜ CS-036 |
| W14-030 (VSA Fabrication) | ✅ Logged | ⬜ PENDING #SessionSummary | ⬜ CS-036 |
| W14-031 (Truncated Response) | ✅ Logged | ⬜ PENDING #SessionSummary | ⬜ CS-036 |
→ Reference: GTM_2026-W14_2037 / GTM_2026-W14_2042 / GTM_2026-W14_2047
→ Lesson: User Data Integrity + Honest Attribution + No Truncation
→ **Full incident documentation: See CS-036 (COMING SOON)**
```

---

## 📋 Model Change History

| Date | Model | Change | CCC-ID | Approved By |
|------|-------|--------|--------|-------------|
| **2026-03-31 19:21** | Qwen3.5 Plus 2026-02-15 | INT-OG1 Restored (Post-VPN fix) | GTM_2026-W14_2041 | @GTM |
| **2026-03-31 18:04** | Qwen3.6 Plus Preview | INT-OG1 Temporary Switch (502/VPN) | GTM_2026-W14_2034 | @GTM |
| **2026-03-31** | Qwen3.6 Plus Preview (FREE) | Added — authorized for all instances | GTM_2026-W14_2017 | @THY + @GTM |
| **2026-03-30** | Qwen3.5 Plus 2026-02-15 | INT-OG1 primary model | GTM_2026-W14_1034 | @GTM |
| **2026-03-28** | Qwen3.5-397B-A17B | INT-M02 Surge ⚡ primary | GTM_2026-W14_1056 | @GTM |
| **2026-03-27** | Xiaomi MiMo-V2-Pro | INT-M02 MiMo 🧪 primary | GTM_2026-W14_1048 | @GTM |
| **2026-03-25** | Claude Opus 4.6 | INT-P01 Calhoun 🎖️ primary | GTM_2026-W14_1021 | @GTM |

---

## 📋 Related Documents

| Document | Version | #masterCCC | Approval | URL |
|----------|---------|------------|----------|-----|
| **BP-068** | v3.2.3.1 | GTM_2026-W12_725 | GTM_2026-W12_728 ✅ | [GitHub](https://github.com/CCCbotNet/fedarch/blob/main/_GOVERNANCE_/BP-068.md) |
| **L-232** | v3.2.5.1 | GTM_2026-W14_1008 | GTM_2026-W14_1015 ✅ | [GitHub](https://github.com/CCCbotNet/fedarch/blob/main/_GOVERNANCE_/L-232.md) |
| **CS-036** | ⬜ DRAFT | GTM_2026-W14_2043 | ⬜ COMING SOON | ⬜ PENDING |
| **PRJ-044** | v3.2.5.1-r11 | GTM_2026-W14_1082 | GTM_2026-W14_1108 ✅ | [GitHub](https://github.com/CCCbotNet/fedarch/blob/main/_PROJECTS_/PRJ-044.md) |
| **INT-OG1_Instance-Plan** | v3.2.5.1 | GTM_2026-W14_1048 | GTM_2026-W14_1048 ✅ | [GitHub](https://github.com/CCCbotNet/fedarch/blob/main/_INSTANCE_/INT-OG1_Instance-Plan.md) |
| **SharedKernel** | v3.2.2.1 | GTM_2026-W11_118 | GTM_2026-W11_139 ✅ | [GitHub](https://github.com/CCCbotNet/fedarch/blob/main/_SYS_/SharedKernel.md) |
| **GOVERNANCE-REGISTRY** | v3.2.2.1 | GTM_2026-W11_541 | GTM_2026-W11_606 ✅ | [GitHub](https://github.com/CCCbotNet/fedarch/blob/main/_SYS_/GOVERNANCE-REGISTRY.md) |
| **SECURITY-REGISTRY** | ⬜ DRAFT | GTM_2026-W14_2028 | ⬜ PENDING | ⬜ PENDING |

---

## 📋 PRJ-044 Alignment

### Gradient™ Migration Model Mapping

| Current (OpenRouter) | Target (Gradient™) | Traffic % |
|----------------------|--------------------|:---------:|
| Qwen3.5 Plus 2026-02-15 | Gradient™ Dedicated | 80% |
| Qwen3.5 Plus 2026-02-15 | Gradient™ Serverless | 15% |
| Qwen3.5 Plus 2026-02-15 | GPU Droplets (vLLM) | 5% |

### PRJ-044 Timeline

| Phase | Timeline | Model Impact |
|-------|----------|--------------|
| Phase 1: Assessment | W14 D2-D3 | ✅ COMPLETE |
| Phase 2: Gradient™ Setup | W14 D4-W15 D1 | ⬜ PENDING |
| Phase 3: Model Deploy | W15 D2-D4 | ⬜ PENDING |
| Phase 4: Cutover | W15 D5 | ⬜ PENDING |

---

## 📋 FOSS Alignment Metrics

### Current Ecosystem FOSS Status

| Metric | Value |
|--------|-------|
| **Total Instances** | 8 |
| **HIGH FOSS Alignment** (🟢) | 6/8 (75%) ✅ |
| **MEDIUM FOSS Alignment** (🟡) | 0/8 (0%) |
| **LOW FOSS Alignment** (🟠) | 2/8 (25%) — Calhoun + INT-P06 |

### FOSS Alignment Ratings

| Rating | Meaning | Criteria |
|--------|---------|----------|
| 🟢 **HIGH** | Open weights + permissive license | Weights public (Apache 2.0, MIT) |
| 🟡 **MEDIUM** | Infrastructure only | Self-hostable |
| 🟠 **LOW** | Proprietary | API-only, no weights |

### Target Goal

| Goal | Current | Target | Gap |
|------|:-------:|:------:|:---:|
| **FOSS Alignment** | 75% | 90%+ | 15% |
| **Timeline** | — | W22 (2026-05-31) | ~8 weeks |

---

## 📋 Version History

| Version | Date | #masterCCC | Approval | Changes |
|---------|------|------------|----------|---------|
| v3.2.5.1-r9 | 2026-W14 D2 | GTM_2026-W14_2049 | ✅ **R-011 APPROVED** | **L-233 REFERENCES REMOVED** (created by Qwen3.6 Plus Preview without governance process); L-233 will be documented within CS-036 instead; All #TriMETA VSA results retained (Calhoun 127/127, Surge 64/64, MiMo 10/10); @GTM Explicit Approval maintained (GTM_2026-W14_2048); INT-OG1 model confirmed as Qwen3.5 Plus 2026-02-15; Filename → `MODEL-REGISTRY.md`; Key Rotation → `SECURITY-REGISTRY.md`; FOSS columns added (§2, §3, §11); BP-068 & L-232 aligned; Single-row Version History enforced; W14-025 (RAG verified) + W14-029/030/031 (⬜ PENDING #SessionSummary → CS-036) documented in §7; #TriMETA ACTUAL dispatch executed (GTM_2026-W14_2044) — NO FABRICATION; RAG Readiness: ✅ OPTIMIZED (~18KB, 12 H2 sections). |

---

## 📝 REVISION LOG (AI WORKFLOW)
`r1 (Init) → r2 (FOSS) → r3 (Structure/Key Rotation) → r4 (Context/VPN) → r5 (Format/VH Row) → r6 (Integrity/1M Fix) → r7 (VSA Fabrication Removed) → r8 (CS-036 Reference + ACTUAL #TriMETA VSA) → r9 (L-233 Removed - No Governance Process)`

---

# 📊 #TriMETA VSA RESULTS — APPENDIX

| Agent | Instance | VSA Type | Score | Max | % | Status | Notes |
|-------|----------|----------|:-----:|:---:|:---:|:------:|-------|
| **Calhoun 🎖️** | INT-P01:tools | Governance | **127** | 127 | **100%** | ✅ PASS | 0 findings, all 4 prior LOWs resolved |
| **Surge ⚡** | INT-M02:tools-qwen | Technical | **64** | 64 | **100%** | ✅ PASS | W14-031 self-reported + corrected |
| **MiMo 🧪** | INT-M02:tools-mimo | Logic | **10** | 10 | **100%** | ✅ PASS | §7 references corrected to CS-036 pending |
| **COMBINED** | — | — | **201** | 201 | **100%** | ✅ **UNANIMOUS** | **GH PUSH READY** |

---

#FlowsBros #FedArch #WeOwnSeason003 #LLMmodel #Registry
#Qwen3-6 #FREE #OpenRouter #GradientAI #PRJ044
#BP-068 #L232 #SecurityFirst
#TriMETA #OptionA #Unanimous #100Percent
#YonksPromptingAcademy #WeCelebrateOurWins #FOSSFirst

🤖 **MODEL-REGISTRY.md v3.2.5.1-r9 — R-011 APPROVED + #TriMETA VSA INTEGRATED.** 
**L-233 REMOVED** (created without governance process - will be in CS-036). 
**Calhoun 127/127 (100%) + Surge 64/64 (100%) + MiMo 10/10 (100%).** 
**@GTM Explicit Approval: GTM_2026-W14_2048.** 
**W14-025 RAG verified. W14-029/030/031 → CS-036 (COMING SOON).** 
**GH PUSH READY.** 🔥🫡

---

## ✅ #BetterUnderstanding — L-233 REMOVED (No Governance Process)

| Field | Value |
|-------|-------|
| **Correction Time** | 20:10 MDT · Tue 31 Mar 2026 |
| **CCC-ID** | **GTM_2026-W14_2049** |
| **Issue** | L-233 created by Qwen3.6 Plus Preview WITHOUT R-011 / #TriMETA VSA |
| **Governance Violation** | Law created without proper approval process |
| **Resolution** | Remove ALL L-233 references from MODEL-REGISTRY.md |
| **Future Reference** | L-233 will be documented within CS-036 (proper case study context) |

> **This is excellent governance catch.** A "Law" cannot be created by an AI during an incident without going through R-011 approval + #TriMETA VSA. CS-036 will document what happened, and L-233 can be properly proposed through governance channels if needed.

---

♾️ WeOwnNet 🌐 ● 🏡 Real Estate and 🤝 cooperative ownership for everyone ● An 🤗 inclusive community, by 👥 invitation only.
---
<!-- CONTENT-HASH-BOUNDARY -->
### ✅ BP-075 CANONICAL HASH — 🛡️ FINALIZE [re-minted by 🛡️ CCC-MAIT-S004-GTM-GLM @ 2026-09-11 22:11 MDT / 2026-09-12T04:11Z · W37 D5 | 11/11 REVIEWS IN (R1 6/6 · R2 5/5 · 🪽 Hermes) — 1 MUST-FIX folded: L52 → GTM_2026-W37_5021 | stage-only, never push · pending @GTM R-011 + @GTM:ADMIN push (#127 NOT charged) | #HITL by @yonks @ 12Sep2026 19:10 EDT]
Content-SHA256: 663aaefa235ec87045bfaad0887aaa52183ebf78f25e4b35f066be96a28931ef
FEDARCH-CANARY: 663aaefa
CHARACTERS: 40782
WORDS: 6082
LINES: 666

---

**♾️ WeOwnNet 🌐 [DUNA][.OCA] Ecosystem** ● 🏡 Real Estate and 🤝 cooperative ownership for everyone ● An 🤗 inclusive community, *by 👥 invitation only*.