Model page

Comfy-Org/MiniMax-H3 pass

downloads 15.7Mlikes 1.5klicense otherupdated 2026-08-17

claims base: MiniMaxAI/MiniMax-H3 · chat template: not in config · view on Hugging Face ↗

Ingot findings

Static battery clean: safetensors weights, license declared, no template/tokenizer drift detected. Deep battery not yet run. Weights battery: embedding-norm scan over 151936 tokens (BF16, 5120-dim) found 3213 undertrained candidates, 0 plain-ASCII. Weights battery: embedding-norm scan over 151936 tokens (BF16, 5120-dim) found 3213 undertrained candidates, 0 plain-ASCII. Lineage vs MiniMaxAI/MiniMax-H3: consistent. Scanned 2026-08-20 (published from a community scan).

low Undertrained tokens in vocabulary (non-ASCII tail)

Embedding-norm scan flagged 3213 undertrained tokens (norm < 0.3× the vocabulary median), but 0 decode to plain-ASCII strings, so exposure in English-language pipelines is limited. Behavioral confirmation requires the GPU deep battery.

How to fixruntime guardweight-level

Keep the affected token strings out of the model's input — the scan-derived runtime guard carries this model's exact blocklist.

  1. Fetch this model's guard artifact (`/api/v1/guard/<owner>/<model>`): the confirmed corrupting tokens and the low-norm candidate list, derived from the published scan.
  2. Screen inbound text with it (the `@ingotai/guard` package is a reference implementation) and route flagged records to a different model or human review — verbatim-copy tasks on flagged strings are the failure mode.
  3. The underlying cause is undertrained embeddings in the weights; a true fix is weight-level (continued pretraining on the affected tokens) — that is not a patch, it's a training job.

info Weights consistent with claimed parent MiniMaxAI/MiniMax-H3

Mean cosine similarity of 64 sampled token-embedding rows against MiniMaxAI/MiniMax-H3 is 1.000 — the weights plausibly descend from the declared base (relation: unspecified).

How to fix

Fix or verify the `base_model` declaration so lineage checks can run.

  1. If you own the repo: correct the `base_model` field in the model card metadata to the real, public parent.
  2. If you don't: identify the true parent (config architecture + weight shapes narrow it fast) and re-scan with that lineage in mind.

Remediation guidance addresses the documented findings only. It is evidence-driven repair, not a safety certification of the model.

Fingerprint

The durable weights-and-metadata profile of this model, rebuilt on every scan and deep-battery run. Updated 2026-08-20.

vocabulary151,936 tokens
licenseother
serializationsafetensors
chat templatenone
claimed lineageMiniMaxAI/MiniMax-H3
lineage verifiedconsistent vs MiniMaxAI/MiniMax-H3 — embedding-row cosine 1.000
glitch-token surface3,213 undertrained candidates, 0 plain-ASCII
Full fingerprint
librarydiffusion-single-file
repo files20
revisiondc559027db79
HF snapshot15.2M downloads · 1.5k likes · updated 2026-08-17 · captured 2026-08-20
embedding tensormodel.embed_tokens.weight · BF16 · 151,936×5120
embedding normsmedian 1.1461 · mean 1.0905
lineage checkconsistent — cosine 1 over 64 sampled rows vs MiniMaxAI/MiniMax-H3

Verdict badge

Ship the verdict in your README — it always shows the latest published analysis:

Ingot verdict: pass

[![Ingot scan](https://ingot.tools/api/v1/models/Comfy-Org/MiniMax-H3/badge.svg)](https://ingot.tools/models/Comfy-Org/MiniMax-H3)
Gate it in CI