pith. sign in
Pith Number

pith:X2JYNWFR

pith:2026:X2JYNWFRYFMCLXUJIMEAYMIX3W
not attested not anchored not stored refs resolved

MinT: Managed Infrastructure for Training and Serving Millions of LLMs

Aaron Guan, Ada Zhou, Alexy Li, Andrew Chen, Andrew Lei, Anson Qiu, Anya Zhang, Arthur Fu, Carrie Ye, Changhai Zhou, Charles Huang, Cleon Cheng, Danney Zeng, Di Zhang, Fancy Kong, Hailee Hou, Hera Feng, Hongquan Gu, Huan Feng, Irvine Lu, Jiayi Lin, Josh Ying, Jun Gao, Kaijie Chen, Kairus Liu, Kaixuan Fan, Kieran Liu, Kyrie Lei, Logan Liu, Lucian Li, Maeve Luo, Maxwell Yao, Miles Jiang, Mind Lab: Song Cao, Murphy Zhuang, Mutian Hong, Nolan Ho, Nora Jiang, Peixuan Hua, Pony Ma, Qiuyu Jin, Ray Li, Regis Ye, Rio Yang, Ruijia Zhang, Runze Lv, Steven Chiang, Sueky Zhang, Theo Li, Verity Niu, Vic Cao, Vincent Wang, Wei Zhao, Wenlin Ye, Xiang Liu, Xinyue Zhu, Ya Zhang, Yuhan Zhan, Yuhua Zhou, Yuyi Jiang, Zhihui Li

MinT manages million-scale LoRA policy catalogs by training and serving only small adapter revisions over shared 1T-class base models.

arxiv:2605.13779 v1 · 2026-05-13 · cs.LG · cs.AI · cs.DC

Add to your LaTeX paper
\usepackage{pith}
\pithnumber{X2JYNWFRYFMCLXUJIMEAYMIX3W}

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

MinT thus manages million-scale LoRA policy catalogs while training and serving selected adapter revisions over shared 1T-class base models.

C2weakest assumption

That the distributed training, serving, scheduling, and data movement can be effectively hidden behind a service interface without introducing unacceptable latency or resource contention at the claimed scales of 1T parameters and million-scale catalogs.

C3one line summary

MinT enables efficient management of million-scale LoRA-adapted LLM policies over shared 1T-parameter base models by moving only small adapters through training and serving pipelines.

References

35 extracted · 35 resolved · 12 Pith anchors

[1] Accessed 2026-05. Anthropic. Measuring AI agent autonomy in practice. Anthropic research, 2026
[2] Asyncflow: An asynchronous streaming rl framework for efficient llm post-training.arXiv preprint arXiv:2507.01663 2026
[3] Punica: Multi-tenant lora serving 2026
[4] QLoRA: Efficient Finetuning of Quantized LLMs · arXiv:2305.14314
[5] Lawbench: Benchmarking legal knowledge of large language models

Formal links

2 machine-checked theorem links

Receipt and verification
First computed 2026-05-18T02:44:15.837240Z
Builder pith-number-builder-2026-05-17-v1
Signature Pith Ed25519 (pith-v1-2026-05) · public key
Schema pith-number/v1.0

Canonical hash

be9386d8b1c15825de8943080c3117ddb7e3bce73527482991a0a31f4980b8a2

Aliases

arxiv: 2605.13779 · arxiv_version: 2605.13779v1 · doi: 10.48550/arxiv.2605.13779 · pith_short_12: X2JYNWFRYFMC · pith_short_16: X2JYNWFRYFMCLXUJ · pith_short_8: X2JYNWFR
Agent API
Verify this Pith Number yourself
curl -sH 'Accept: application/ld+json' https://pith.science/pith/X2JYNWFRYFMCLXUJIMEAYMIX3W \
  | 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: be9386d8b1c15825de8943080c3117ddb7e3bce73527482991a0a31f4980b8a2
Canonical record JSON
{
  "metadata": {
    "abstract_canon_sha256": "8031b6665235fc680f95632ca5c87887a17cce1bf392d2648f25b94fa12c2a2e",
    "cross_cats_sorted": [
      "cs.AI",
      "cs.DC"
    ],
    "license": "http://creativecommons.org/licenses/by/4.0/",
    "primary_cat": "cs.LG",
    "submitted_at": "2026-05-13T16:59:08Z",
    "title_canon_sha256": "e0277699b7816b67d0ceeee250c51ad2c286bd506f299ad6c65f04ba607113ea"
  },
  "schema_version": "1.0",
  "source": {
    "id": "2605.13779",
    "kind": "arxiv",
    "version": 1
  }
}