WikiWayne
Local AIWikiBenchGamesGrok BotAI ToolsTech NewsAboutBlogContact

As an Amazon Associate I earn from qualifying purchases. This site contains affiliate links.

WikiWayne

Independent notes on local AI, real hardware, Grok, and the games that come out of that stack.

Categories

  • Local AI Hub
  • Local AI Glossary
  • Local AI
  • AI Tools
  • Tech News

Quick Links

  • About Wayne
  • Games
  • Grok Bot
  • Contact
  • Links
  • ADHD MaxxxingSponsored
  • WikiBench
  • Methodology
  • Editorial Standards
  • Disclosures
  • Privacy Policy
  • Sitemap

Follow on X

Hardware, local models, Grok, and what I actually ship.

Follow @wikiwayne
WikiWayne© 2026
PrivacyWikiBenchMethodologyEditorialDisclosuresTermsSitemap

Disclosure: As an Amazon Associate I earn from qualifying purchases. This site contains affiliate links.

Home/Local AI/Nemotron Puzzle Fits 128GB on Paper. Loader Fix Landed. Not a GX10 Sit.
Back to Blog
Nemotron Puzzle Fits 128GB on Paper. Loader Fix Landed. Not a GX10 Sit.
Local AI

Nemotron Puzzle Fits 128GB on Paper. Loader Fix Landed. Not a GX10 Sit.

Published: September 4, 2026

Official BF16 is 156,613,878,652 bytes; FP8 is 83,064,072,032; NVFP4 is 53,537,521,983. FP8 and NVFP4 fit 128GB on disk. The official card calls it 75.3B total and 9.3B active.

Key takeaways

  • Official BF16 is 156,613,878,652 bytes; FP8 is 83,064,072,032; NVFP4 is 53,537,521,983. FP8 and NVFP4 fit 128GB on disk. The official card calls it 75.3B total and 9.3B active.
  • Architecture is nemotron_h_puzzle: 88 blocks (40 Mamba2, 8 attention, 40 MoE). MoE is heterogeneous by layer. supports_mtp_export=False — skip MTP/NEXTN quants.
  • llama.cpp PR 25444 merged the architecture. The expert-layer loader failure needed PR 28323 (n_expert_used_max), merged 2026-09-04T04:36:51Z. First tagged nightly with that fix is b10796 (published 2026-09-04T05:31:09Z). Latest checked: b10798. RemySkye Q4_K_M 51,601,807,488 / Q8_0 83,453,368,448 unchanged. I have not loaded this on the GX10. Not a sit.
3 min read
local-ai, hardware, open-weights
Wayne Lowry, WikiWayne author
Wayne Lowry

Local LLMs on NVIDIA Spark / ASUS GX10

The headline is deliberately split in three. The puzzle fits 128GB on paper. The expert-layer loader fix landed in llama.cpp. I still have not sat it on the GX10.

The official sizes are the first receipt: BF16 is 156,613,878,652 bytes; FP8 is 83,064,072,032; NVFP4 is 53,537,521,983. FP8 and NVFP4 fit 128GB as an on-disk claim. BF16 does not. None of those numbers is leftover memory, context headroom, or a speed result.

The official card says 75.3B total / 9.3B active. License field: other. The license is the OpenMDW License Agreement 1.1. I am not laundering that into Apache or MIT.

The model type is nemotron_h_puzzle. There are 88 blocks: 40 Mamba2, 8 attention, and 40 MoE. The MoE layers are heterogeneous per layer, which matters here. supports_mtp_export=False. There are no MTP/NEXTN quants. Skip those. This is not a plain dense checkpoint with a familiar loader path.

The llama.cpp receipt

llama.cpp PR 25444 merged at 2026-09-03T06:53:09Z, commit c61b98b875eaa5e654a3f5c73b34c310d2c6ab4c. That got the architecture into the tree. It was not the same thing as a successful load. The loader errors that followed were:

model has expert layers but no expert layers are used
GGML_ASSERT(n_ids_used > 0) failed

PR 28323 (n_expert_used_max) closed that gap. It merged at 2026-09-04T04:36:51Z, merge commit 9a4843cf2f1a3fc8e39f8148e92ee6bfe18e2db6. The first tagged nightly carrying that fix is b10796, published 2026-09-04T05:31:09Z (target commit matches the merge). Latest nightly I checked is b10798, published 2026-09-04T15:03:27Z.

Sponsored

ADHD IS MY SUPER POWER. Wear it proud. Shop adhdmaxxxing.com.

Two nearby receipts are real but do not replace this one. b10778 / PR 27483 reduces load peak RAM through per-tensor buffers. b10782 / PR 28198 adds concurrent streams per split for multi-GPU CUDA. Neither was the expert-count fix.

A merged loader fix is still not my GX10 sit. I have not loaded Nemotron Puzzle on this box, and I am not inventing leftover unified memory or tokens per second.

Community GGUFs are a separate line

RemySkye's listed community quants are unchanged in this check: Q4_K_M 51,601,807,488, Q8_0 83,453,368,448. Those are community conversions, not official NVIDIA artifacts. The sizes support an on-disk fit statement for those individual files. They do not supply a GX10 result.

Danbev manually converted and tested Q8_0 and Q4_K_M on a DGX Spark. There is no tok/s in this note and no logits verification. That is useful conversion evidence, not my GX10 bench. I am not quoting it as one.

So the status is narrow: official FP8, official NVFP4, and the listed community quants fit 128GB on disk. The expert-layer loader fix is in llama.cpp as of b10796. I have not sat the puzzle on the GX10. This is not a GX10 sit.

I run Grok when it earns it, and a GX10 when I want the weights in the room. This one is a size claim plus a loader ticket that closed. When I have actual GX10 receipts, that will be a different article.

  • NVIDIA Nemotron models
  • OpenMDW License Agreement 1.1
  • llama.cpp PR 25444
  • llama.cpp PR 28323
  • llama.cpp PR 27483
  • llama.cpp PR 28198
  • llama.cpp b10796
  • llama.cpp b10798
  • llama.cpp commit c61b98b875eaa5e654a3f5c73b34c310d2c6ab4c
  • RemySkye community GGUFs

Frequently asked questions

The official FP8 file is 83,064,072,032 bytes and the official NVFP4 file is 53,537,521,983 bytes, so both fit as an on-disk claim. The official BF16 file is 156,613,878,652 bytes and does not. Fit is not a run.

The expert-layer loader fix is in the tree. PR 28323 merged at 2026-09-04T04:36:51Z and first showed up in tagged nightly b10796. I have not run a load myself on the GX10, so I am not claiming a successful sit from this card.

No. danbev manually converted and tested Q8_0 and Q4_K_M on a DGX Spark, without tok/s or logits verification. That is not my GX10 bench, and I am not turning it into one.

Sponsored

ADHD MAXXXING. Merch for people with ADHD who wear it proud. Shop adhdmaxxxing.com.
ADHD MAXXXING. Merch for people with ADHD who wear it proud. Shop adhdmaxxxing.com.

Affiliate Disclosure: As an Amazon Associate I earn from qualifying purchases. This site contains affiliate links.

Related Articles

local ai

AngelSlim Hy4 GGUFs Are Real Weights. Not a GX10 Sit.

2 min read

local ai

llama.cpp 0.3.0 Is a Tag, Not a GX10 Bench

2 min read

local ai

Nex-N2.5-mini Fits 128GB. That Is Not a Sit.

4 min read