Skip to content
AI Atlas
ModelActiveOpen weights

KAT-Dev-72B-Exp

Kwaipilothuggingface.co/Kwaipilot/KAT-Dev-72B-Exp

Open in Graph
data quality55

Updated 2 h ago · first seen 12 Sept 2026

model_01M2AS5J1QKG7CEH1JMGH44E73

Overview

Identity

Canonical model
Yesidentity confidence: highOne row per real model release. Artifacts (checkpoints, quantisations, conversions) and folded evaluation variants point here.
Official checkpoints
official_checkpoints = hf_repo identifiers carried by the model itself; artifacts are separate entities pointing here through canonical_id.
Artifacts
1 quantization0 official · 1 third-partySeparate entities (checkpoint · quantization · conversion · packaging) pointing to this model through canonical_id.
Provider deployments
None recorded
API aliases
NoneIdentifiers under which providers and evaluators refer to this model.
Folded evaluation variants
0Effort / thinking variants (…-high, …-non-reasoning) are result configurations of this model, not separate models. Their old URLs redirect here.

Openness

Open weights— weights downloadable under Apache-2.0; commercial use allowed; redistribution allowed; derivatives allowed; 4 dimensions unknown.

Weights downloadable under a permissive or Creative Commons licence allowing commercial use; code or data may be missing.

  • Weights

    Yes

  • Inference code

    —

  • Training code

    —

  • Training data

    —

  • Dataset

    —

  • Commercial use

    Yes

  • Redistribution

    Yes

  • Derivatives

    Yes

Licence: Apache License 2.0 (permissive · SPDX Apache-2.0)

dimensions marked null are unknown, not false

Key facts

Release date

Source:Hugging Face Hub (public pages, model cards, papers)T2observed 3 h agomedium

Architecture

Architecture

Source:Hugging Face Hub (public pages, model cards, papers)T2observed 3 h agomedium

Model type

Source:Hugging Face Hub (public pages, model cards, papers)T2observed 3 h agomedium

Parameters

Source:Hugging Face Hub (public pages, model cards, papers)T2observed 3 h agomedium

Weights dtype

Source:Hugging Face Hub (public pages, model cards, papers)T2observed 3 h agomedium

File size

Source:Hugging Face Hub (public pages, model cards, papers)T2observed 3 h agomedium

Library name

Source:Hugging Face Hub (public pages, model cards, papers)T2observed 3 h agomedium

Pipeline tag

Source:Hugging Face Hub (public pages, model cards, papers)T2observed 3 h agomedium

Hugging Face repo

Source:Hugging Face Hub (public pages, model cards, papers)T2observed 3 h agomedium

Capabilities

Modalities

Modalities
text
Input
text
Output
text

Capabilities

  • Tool calling

    Unavailable

  • Structured output

    Unavailable

  • Reasoning

    Unavailable

  • Vision

    Unavailable

  • Audio

    Unavailable

  • Fine-tuning available

    Unavailable

No capability flags have been observed from a source yet — we do not infer them.

No structured attributes yet.

Hardware fit37

Estimated

23 of 37 device × quantization combinations fit.

Run locally: your machine →
Estimated hardware fit
HardwareQuantizationDevice memoryEst. memoryFits
Apple M3 Ultra4bit—42.3 GB est.Yes
Mac Studio (Apple M5 Ultra)4bit—42.3 GB est.Yes
Apple M2 Ultra4bit—42.3 GB est.Yes
Apple M1 Ultra4bit—42.3 GB est.Yes
Apple M3 Max4bit—42.3 GB est.Yes
Apple M4 Max4bit—42.3 GB est.Yes
Mac Studio (Apple M5 Max)4bit—42.3 GB est.Yes
MacBook Pro (Apple M5 Max)4bit—42.3 GB est.Yes
Apple M2 Max4bit—42.3 GB est.Yes
Apple M1 Max4bit—42.3 GB est.Yes
Apple M4 Pro4bit—42.3 GB est.Yes
Mac mini (Apple M5 Pro)4bit—42.3 GB est.Yes
MacBook Pro (Apple M5 Pro)4bit—42.3 GB est.Yes
NVIDIA A100 80GB4bit80 GB42.3 GB est.Yes
NVIDIA H100 SXM4bit80 GB42.3 GB est.Yes
NVIDIA H100 NVL4bit94 GB42.3 GB est.Yes
NVIDIA DGX Spark4bit128 GB42.3 GB est.Yes
NVIDIA H2004bit141 GB42.3 GB est.Yes
NVIDIA H200 NVL4bit141 GB42.3 GB est.Yes
NVIDIA B2004bit180 GB42.3 GB est.Yes
AMD Instinct MI300X4bit192 GB42.3 GB est.Yes
AMD Instinct MI325X4bit256 GB42.3 GB est.Yes
NVIDIA DGX B2004bit1,440 GB42.3 GB est.Yes
Apple M3 Pro4bit—42.3 GB est.No
Apple M1 Pro4bit—42.3 GB est.No
Apple M2 Pro4bit—42.3 GB est.No
Apple M44bit—42.3 GB est.No
iMac (Apple M4)4bit—42.3 GB est.No
Mac mini (Apple M6)4bit—42.3 GB est.No
MacBook Air (Apple M5)4bit—42.3 GB est.No
MacBook Pro (Apple M5)4bit—42.3 GB est.No
Apple M24bit—42.3 GB est.No
Apple M34bit—42.3 GB est.No
Apple M14bit—42.3 GB est.No
NVIDIA GeForce RTX 30904bit24 GB42.3 GB est.No
NVIDIA GeForce RTX 40904bit24 GB42.3 GB est.No
NVIDIA GeForce RTX 50904bit32 GB42.3 GB est.No
Assumptions (7)
  • Estimated, not measured: weights = parameters × bytes/param × 1.15 runtime overhead (or the observed artifact file size when one is recorded).
  • bytes/param: 4bit = 0.5, 8bit = 1.0, fp16 = 2.0 (uniform quantization, no per-layer exceptions).
  • KV cache: 2 × layers × kv_heads × head_dim × 2 bytes × context × batch when the architecture is known; otherwise 0.5 GB per 8 192 tokens (× batch), independent of architecture (GQA/MLA models need less).
  • A model 'fits' when the estimate is at most the device memory minus 2 GB reserved for the OS and framework.
  • Mixture-of-experts models are estimated on total parameters (all experts must be resident); active parameters are ignored.
  • Device memory uses the largest configuration when several are listed (e.g. Apple silicon tiers).
  • Multi-GPU: device memories are summed; interconnect bandwidth, tensor-parallel replication and pipeline bubbles are not modelled.

Lineage

Open in Graph →
Explicit derived_from / fine_tuned_from / distilled_from relations stated by sources; artifacts collapsed by kind.
DESCENDANTS 0 · ARTIFACTS1 quantizationartifacts · collapsed1 quantization — artifacts · collapsedKAT-Dev-72B-Exp72.7B params · this modelKAT-Dev-72B-Exp — 72.7B params · this model

    Versions & Artifacts1

    Version history

    Licensefirst observation only

    12 Sept 2026current

    Opennessfirst observation only

    12 Sept 2026current

    Parametersfirst observation only

    12 Sept 2026current

    Each hop is a claim: click a value for its source, tier and observation time. Nothing is overwritten — a new observation closes the previous claim.

    Artifacts 1

    quantization 1

    Timeline1

    Full timeline →

    Change history

    Temporal, append-only claims: a new observation closes the previous claim instead of overwriting it. Rewind the record with the as-of picker.
    1 claims · 1 propertiesShow all properties

    Downloads all timemetric.downloads_all_time1

    Claim history for Downloads all time
    ValueValid from → toStatusSourceConfidenceExtractor
    6,038 → currentcurrentHugging Face Hub (public pages, model cards, papers)T2mediumdeterministic

    Claims are temporal and append-only: a new observation closes the previous claim (valid_to) instead of overwriting it. Conflicting claims from different sources are kept side by side and flagged — never averaged. Methodology →

    Provenance

    Attributed facts

    25

    Source tiers

    T225

    Freshest observation

    3 h ago

    Conflicts

    None

    Source documents 3

    Source documents
    SourceDocumentTypeTierLast observedSnapshots
    Hugging Face Hub (public pages, model cards, papers)huggingface.co/Kwaipilot/KAT-Dev-72B-Exp/raw/main/README.md model_cardT2· Quality secondary2 h ago1
    Hugging Face Hub (public pages, model cards, papers)huggingface.co/Kwaipilot/KAT-Dev-72B-Exp model_pageT2· Quality secondary3 h ago1
    Hugging Face Hub (public pages, model cards, papers)huggingface.co/models?author=Kwaipilot&p=0&sort=downloads listingT2· Quality secondary3 h ago2

    Tier 1 = official/primary, 2 = quality secondary, 3 = community, 4 = unverified. Every snapshot is archived; see all sources and the methodology.

    Data quality (55/100) measures how well AI Atlas knows this entity — completeness, primary-source ratio, freshness, conflicts — never how good the model is. Methodology →