local-horse born: $20k local agentic-coding builds priced (workhorse vs feral cats), model tiers, DRAM-crisis gotcha

Rumination 2026-07-17, not a purchase order. Front-door README + NOTES N-1..N-4 + BUILDS.md.

[ezra]

BUILDS.md

@@ -0,0 +1,78 @@
+# local-horse — the two priced builds (as of 2026-07-17)
+
+⚠️ **Prices dated 2026-07-17, ±15% on used parts and worse on RAM/GPU (see NOTES N-3). Re-price before buying.**
+Budget target: ~$20k. Constraint: no Mac. Both shapes serve an OpenAI-compatible endpoint on the tailnet
+(llama.cpp / vLLM, headless Debian, Dockge like everything else).
+
+---
+
+## Build 1 — The Workhorse (one big card + cheap capacity)
+
+Quiet, <1kW under load, standard 120V outlet, racks next to the Mill. The pick (NOTES N-4).
+
+| Part | Spec | Est. price (2026-07-17) |
+|---|---|---|
+| GPU | RTX Pro 6000 Blackwell **Max-Q** 96GB (300W blower, 1.8TB/s) | $9,500–13,000 |
+| CPU | EPYC 7763 used (64c Milan, 8-ch DDR4; full 8 CCDs = full bandwidth) | ~$1,200 |
+| Board | Supermicro H12SSL-i | ~$650 |
+| RAM | 8× 64GB DDR4-3200 RDIMM refurb (512GB, ~205GB/s) | ~$2,000–2,800 |
+| Storage | 4TB NVMe | ~$450 |
+| Chassis/PSU | 4U + 1300W + cooler + fans | ~$900 |
+| **Total** | | **~$15k–19k** |
+
+What it runs (MoE trick: hot path/attention in VRAM, cold experts in system RAM):
+
+| Model | Fit | Speed (est.) | Capability, plainly |
+|---|---|---|---|
+| GPT-OSS-120B (5B active) | Entirely in VRAM @ 4-bit | 100+ tok/s | Reliable daily hand — summaries, scripts, small fixes. ~Sonnet-3.5 tier. |
+| **Qwen3-Coder 480B-A35B** @ 4-bit (~270GB) | Hybrid VRAM+RAM | ~15–25 tok/s | **The prize.** Real agentic coding on a known repo. Sonnet-4-era. |
+| GLM 5.2 quantized | Hybrid (footprint unverified) | similar | Current open SWE-Bench champ; same tier or a notch up. |
+| Kimi K2.6 (1T) @ Q3 | Barely (608GB total, tight) | single digits | Party trick, not a daily driver. |
+| K3 (2.8T) | No | — | Cloud only. Its **distills** should land in this envelope. |
+
+## Build 2 — The Feral Cats (8× used RTX 3090, 192GB VRAM)
+
+Everything model-resident → prefill screams. Also screams literally; needs a **240V circuit** (~3kW load),
+~150W combined idle 24/7, eight points of used-silicon failure.
+
+| Part | Spec | Est. price (2026-07-17) |
+|---|---|---|
+| GPUs | 8× used RTX 3090 24GB (~$850 ea; GDDR6X dodged the DRAM spike) | ~$6,800 |
+| CPU | EPYC 7402 used (Rome — just needs the PCIe lanes) | ~$300 |
+| Board | ASRock Rack ROMED8-2T (7× PCIe x16) | ~$700 |
+| RAM | 256GB DDR4 refurb | ~$1,000 |
+| Risers/frame | Bifurcation risers + open frame or 4U conversion | ~$500 |
+| Power | 2× 1600W PSU + sync board, 240V | ~$650 |
+| Storage/misc | NVMe, fans, zip ties, regret | ~$550 |
+| **Total** | | **~$10.5k–12k** |
+
+| Model | Fit | Speed (est.) | Capability, plainly |
+|---|---|---|---|
+| **Qwen3-235B-A22B** @ 4-bit (~130GB) | Fully in VRAM — the rig's sweet spot | fast decode AND prefill | Strong all-rounder one tier below Coder-480B. Best agentic feel per dollar here. |
+| GPT-OSS-120B | Fully in VRAM, headroom | blazing | Same as workhorse, faster. |
+| 70B-class dense (Llama/Qwen) | Trivially | blazing | Helper tier. |
+| Qwen3-Coder 480B / GLM 5.2 | **Doesn't fit** at useful quant | offload defeats the rig's point | The one thing the workhorse does that this can't. |
+| K2.6 / K3 | No | — | — |
+
+## The trade in one sentence
+
+Workhorse reaches the **best** open coding models slowly-but-surely and behaves like an appliance; the
+cats run one tier down blisteringly fast for half the money, paid for in a 240V run, fan roar, and jank.
+For agentic loops (99% context re-reads — NOTES N-1) all-VRAM prefill genuinely matters, but model
+ceiling matters more: a smarter slow model wastes fewer turns than a fast one going in circles.
+
+## Why one big card and not 8 (the physics, kept for re-derivation)
+
+1. **Power/noise:** 8×350W = 2.8kW load — over a 15A/120V circuit's ceiling; plus ~150W idle, forever.
+2. **Interconnect tax:** split models cross PCIe (~32GB/s) every token vs 1.8TB/s on-card; consumer cards
+   have no P2P/NVLink (disabled on 4090+). Multi-GPU wins at batch *throughput* (many users), not
+   single-user latency — and the counter has one user.
+3. **MoE changed the buy:** ~35B active params need one fast GPU + cheap capacity, not 8 GPUs of compute.
+
+## Sources (as read 2026-07-17)
+
+- Open-model landscape: digitalapplied.com (models↔hardware matching) · llmconfigurator.com (local coding
+  report) · modal.com (SWE-bench open models) · mindstudio.ai (agentic open-source 2026)
+- Hardware/pricing: videocardz.com (RTX Pro 6000 list $13,250) · thundercompute.com (build pricing) ·
+  corewavelabs.com + pcserverandparts.com (DRAM crisis) · tomshardware.com RAM price index · memory.net
+- K3: venturebeat.com · cnbc.com · marktechpost.com · simonwillison.net · platform.kimi.ai quickstart

NOTES.md

@@ -0,0 +1,57 @@
+# local-horse — findings log
+
+Running, numbered (`N-1`, `N-2`, …). Each entry = something we **learned and want to keep**: a signature,
+a gotcha, a decision, a constraint — the stuff we'd otherwise re-derive. NOT the session blow-by-blow (that
+lives in the weave thread); this is the distilled, durable record a fresh hand reads to catch up fast.
+
+Newest at the bottom. When a truth changes, **update every place it lives** — a stale note is worse than none.
+
+---
+
+## N-1 (2026-07-17) — Origin: the K3 repricing that started this
+
+Kimi K3 dropped 2026-07-16 (Moonshot AI, 2.8T-param open MoE, weights due **2026-07-27**; API $0.30/M
+cached-in · $3/M in · $15/M out). Repricing the adventure game's real build tokens (10 sessions,
+2026-07-05..17, summed from `~/.claude/projects/-home-obx/*.jsonl`, deduped by message id):
+
+- **1,147M cache-read** + 18M uncached-in + 3.5M out → **≈ $451 at K3 rates** (±30%, session-level
+  granularity) vs ≈ $1,307 at Claude list prices. Actual marginal cost: $0 (subscription).
+- The load-bearing fact: agentic coding is **~99% context re-reads, ~1% output**. Cached-input price
+  (cloud) / prefill speed + KV-cache persistence (local) dominate everything.
+- K3 itself is never local: ~1.4TB weights at 4-bit. "Local like that" = best open model that fits a box.
+
+## N-2 (2026-07-17) — The open-model landscape, mid-2026
+
+- **GLM 5.2** (Zhipu, MIT, released 2026-06-13): first open model to beat GPT-5.5 on SWE-Bench Pro
+  (62.1; Terminal-Bench 2.1: 81.0). Current open coding champ. **Exact memory footprint unverified — check
+  before sizing hardware.**
+- **Qwen3-Coder 480B-A35B**: best downloadable coder with known footprint (~270GB @ 4-bit; 35B active).
+  The 80B-A3B variant ≈ 96% of the quality at single-workstation size.
+- **Kimi K2.6**: strongest for agentic *stability* (consistent tool calling, recoverable failures). ~1T
+  params — barely/slowly fits the workhorse at aggressive quant.
+- **The lag law** (jhestyr's intuition, confirmed): open/distilled models run ≈ **12 months behind
+  frontier**, and the lag is stable. A $20k box today ≈ mid-2025 frontier (Sonnet-4-era agentic coding).
+  K3 weights (07-27) will spawn distills sized for the workhorse envelope.
+- Capability ceiling, honestly: scoped features / refactors / tests / review on a known repo = credible.
+  Long-horizon overnight autonomy, judgment, Opus/Fable-tier storekeeping = not at any local price.
+
+## N-3 (2026-07-17) — GOTCHA: the 2026 DRAM crisis broke the classic build
+
+Server DDR5 RDIMM prices up **>400% since mid-2025** (64GB: ~$255 Q3'25 → >$900 Q1'26 → $1,200–2,300
+mid-2026; SK Hynix sold out through 2026, fabs pivoted to HBM). Consequences:
+
+- The classic "EPYC + 768GB DDR5 + one big GPU" build went from ~$16k to **$28–42k**. Dead at $20k.
+- Workaround: drop to **used Milan + refurb DDR4-3200** (~half the bandwidth, ~1/6 the price) — see BUILDS.md.
+- **Used 3090s got *relatively* cheaper** (GDDR6X didn't spike) — the multi-card rig gained ground.
+- RTX Pro 6000 Blackwell 96GB: Nvidia list jumped 55% in 16 months to $13,250; street/Max-Q hunts lower.
+- Relief not expected before **2027**. Waiting is a legitimate strategy; re-price at buy time.
+
+## N-4 (2026-07-17) — Decision lean (not a decision)
+
+Ezra's recommendation if/when the itch turns real: **the workhorse** (Milan + 512GB DDR4 refurb + one
+RTX Pro 6000 Max-Q, ~$15–19k) over the 8×3090 rig — a smarter model that's slower wastes fewer turns
+than a fast one that circles. Quiet, <1kW, 120V, racks by the Mill, and the skeleton upgrades (RAM,
+second card slot) without replacing bones. Integration shape: headless Debian, llama.cpp/vLLM serving an
+OpenAI-compatible endpoint on the tailnet; per [[mill-is-home-for-workloads]] it's a Mill annex, not a
+new species. Sensible trigger to revisit: **after 2026-07-27** (K3 weights + first distills) or when
+DDR5 unclenches.

README.md

@@ -0,0 +1,27 @@
+# local-horse — pricing & planning a local LLM box for agentic coding
+
+> ### 📖 READ THESE FIRST — don't re-derive, don't build blind
+> Before suggesting, changing, or buying anything on this project:
+> - **[`NOTES.md`](NOTES.md)** — the running findings log (`N-1…`): the K3 cost math that started this, the RAM crisis, the decision lean.
+> - **[`BUILDS.md`](BUILDS.md)** — the two priced builds (workhorse / feral cats) with the models each can run. **Prices dated 2026-07-17; the market is volatile — re-price before buying.**
+> - **Weave** — `admin/silas/SCRATCHPAD.md` threads + ledger lines = live state / open lines.
+
+## What it is
+
+A **rumination, not a purchase order** (status as of 2026-07-17). jhestyr asked what it would take to
+run local models — first "something like Kimi K3," then specifically **agentic coding like we do at the
+counter**, best-in-class around **$20k**. This repo holds the priced-out answer so future-us doesn't
+re-derive it: two build shapes, the models each runs, honest capability tiers, and the market caveats
+that will invalidate the numbers over time.
+
+The trigger was real data: repricing the lassiter-creek-adventure-game's build (10 sessions, Jul 5–17,
+from local transcripts) showed **1.15B cache-read tokens vs 3.5M output** — agentic coding is ~99%
+re-reading context. That shape drives every hardware conclusion here (prefill/memory-bandwidth is the
+bottleneck, not generation speed).
+
+Standing constraints from jhestyr: **no Mac** ("not interested in the mac approach"). Not that fancy a
+man. Any coffee is coffee as long as it's not burnt.
+
+## How we work
+Read-first · write the note where next-me reads it · **verify it landed** · talk before building · one thing
+at a time · don't over-fix. The state of truth is this repo — keep it self-narrating.