Pith Number

pith:5YPLY3O4

pith:2024:5YPLY3O4LHGJZW3YZ6BQJN3OPY

not attested not anchored not stored refs resolved

Pixtral 12B

Albert Q. Jiang, Alexandre Sablayrolles, Am\'elie H\'eliou, Andy Lo, Arthur Mensch, Baptiste Bout, Baptiste Rozi\`ere, Baudouin De Monicault, Devendra Chaplot, Diego Las Casas, Diogo Costa, Emma Bou Hanna, Guillaume Lample, Jessica Chudnovsky, Joachim Studnia, Kartik Khandelwal, Lawrence Stewart, Louis Martin, Lucile Saulnier, Marie Pellat, Nikhil Raghuraman, Patrick von Platen, Paul Jacob, Pavankumar Muddireddy, Pierre Stock, Pravesh Agrawal, Romain Sauvestre, Roman Soletskyi, Sagar Vaze, Sandeep Subramanian, Saurabh Garg, Soham Ghosh, Sophia Yang, Szymon Antoniak, Teven Le Scao, Theophile Gervet, Thibaut Lavril, Thomas Wang, Timoth\'ee Lacroix, Valera Nemychnikova, Wendy Shang, William Marshall

Pixtral-12B outperforms similar and larger open multimodal models by processing images at their native resolution and aspect ratio.

arxiv:2410.07073 v2 · 2024-10-09 · cs.CV · cs.CL

Open paper page JSON Open Graph Bundle Merged state Verified badge What is a Pith Number?

Add to your LaTeX paper

\usepackage{pith}
\pithnumber{5YPLY3O4LHGJZW3YZ6BQJN3OPY}

Prints a linked badge after your title and injects PDF metadata. Compiles on arXiv. Learn more · Embed verified badge

Record completeness

1 Bitcoin timestamp

2 Internet Archive

3 Author claim open · sign in to claim

4 Citations open

5 Replications open

✓ Portable graph bundle live · download bundle · merged state

The bundle contains the canonical record plus signed events. A mirror can host it anywhere and recompute the same current state with the deterministic merge algorithm.

Claims

C1strongest claim

Pixtral-12B substantially outperforms other open models of similar sizes (Llama-3.2 11B & Qwen-2-VL 7B). It also outperforms much larger open models like Llama-3.2 90B while being 7x smaller.

C2weakest assumption

That the reported benchmark scores reflect fair, standardized evaluation without undisclosed differences in training data scale, filtering, or test-set contamination.

C3one line summary

Pixtral-12B is a 12B multimodal LLM with a custom vision encoder that ingests images at native resolution and aspect ratio, achieving leading benchmark results among open models while preserving text capabilities.

References

26 extracted · 26 resolved · 11 Pith anchors

[1] The Claude 3 Model Family: Opus, Sonnet, Haiku 2024

[2] Bavishi, R., Elsen, E., Hawthorne, C., Nye, M., Odena, A., Somani, A., and Ta¸ sırlar, S. (2023). Fuyu-8b: A multimodal architecture for ai agents 2023

[3] Dehghani, M., Mustafa, B., Djolonga, J., Heek, J., Minderer, M., Caron, M., Steiner, A., Puigcerver, J., Geirhos, R., Alabdulmohsin, I. M., et al. (2024). Patch n’pack: Navit, a vision transformer for 2024

[4] Molmo and PixMo: Open Weights and Open Data for State-of-the-Art Vision-Language Models 2024 · arXiv:2409.17146

[5] An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale 2020 · arXiv:2010.11929

Formal links

2 machine-checked theorem links

Cited by

46 papers in Pith

GAVEL: Grounded Caption Error Verification and Localization

Perception, Verdict, and Evolution: Hindsight-Driven Self-Refining Forensics Agent for AI-Generated Image Detection

ESC: Emotional Self-Correction for Reliable Vision-Language Models

Qwen-RobotWorld Technical Report: Unifying Embodied World Modeling through Language-Conditioned Video Generation

Decodable Is Not Grounded: A Vision-Ablation Arbiter for VLM Spatial Reasoning

Receipt and verification

First computed	2026-05-17T23:39:19.791252Z
Builder	pith-number-builder-2026-05-17-v1
Signature	Pith Ed25519 (`pith-v1-2026-05`) · public key
Schema	pith-number/v1.0

Canonical hash

ee1ebc6ddc59cc9cdb78cf8304b76e7e351471bd972f2a658f6e159b7f025e2d

Aliases

arxiv: 2410.07073 · arxiv_version: 2410.07073v2 · doi: 10.48550/arxiv.2410.07073 · pith_short_12: 5YPLY3O4LHGJ · pith_short_16: 5YPLY3O4LHGJZW3Y · pith_short_8: 5YPLY3O4

Agent API

Resolver JSON Graph JSON Events JSON Schema Signing key

Verify this Pith Number yourself

curl -sH 'Accept: application/ld+json' https://pith.science/pith/5YPLY3O4LHGJZW3YZ6BQJN3OPY \
  | jq -c '.canonical_record' \
  | python3 -c "import sys,json,hashlib; b=json.dumps(json.loads(sys.stdin.read()), sort_keys=True, separators=(',',':'), ensure_ascii=False).encode(); print(hashlib.sha256(b).hexdigest())"
# expect: ee1ebc6ddc59cc9cdb78cf8304b76e7e351471bd972f2a658f6e159b7f025e2d

Canonical record JSON

{
  "metadata": {
    "abstract_canon_sha256": "5ccc3122e53fc8ad76ad168b34e3934d70248678eb2a61aad11041f1ce918560",
    "cross_cats_sorted": [
      "cs.CL"
    ],
    "license": "http://creativecommons.org/licenses/by/4.0/",
    "primary_cat": "cs.CV",
    "submitted_at": "2024-10-09T17:16:22Z",
    "title_canon_sha256": "3f972522ad7f9c9b8d3bfef4a659fa191a919883ed15c79909f74eaec56c52da"
  },
  "schema_version": "1.0",
  "source": {
    "id": "2410.07073",
    "kind": "arxiv",
    "version": 2
  }
}