models · moonshot-ai

Kimi K2.6

Moonshot AI Kimi K2.6. 1000B MoE, 262K context, 56 tok/s. Arena 1557. GPQA Diamond 90.5%, HumanEval 86.7%, LiveCodeBench 80.1%. Top open-weight coding model.

Compare Verify

What this registry has checked

Quarantine. We hold the publisher’s own published hash for these weights and have recorded it against this entry. We have not downloaded and hashed the weights ourselves, so the hash is the publisher’s claim, not our measurement.
Weight file SHA-2566b8dcde04bdace13ae7b8552a2d04b40610927a0bbecef9a229b2f5de97a0ce3
Methodhf-published-sha256

What is recorded

Publishermoonshot-ai
OriginRecorded as a covered nation under 10 U.S.C. § 4872(f) — the PRC, Russia, Iran or North Korea. The registry did not record which.
LicenceKimi-Open
Publisher’s own licence statementother
Parameters1,026,879,376,368 (1026.9B)
Context window262K
Modalitytext+image->text
Native precisionI32
Pulls recorded here390,071
Downloads reported upstream539,680

Reported scores

GPQA Diamond91.1
IFBench76
Humanity's Last Exam37.5

Reported by a third party and recorded here, not re-run by us. Source: artificialanalysis.ai/api/v2.

Measured on real hardware

None of these runs are ours. Each figure is the reporter’s, linked to their own evidence, under their own trust label. The number shown is single-stream decode — one user, concurrency 1 — because aggregate throughput across many users is a different quantity and reads as far faster than anyone will see.

Reported to local-ai-registry

From local-ai-registry (MIT), commit 124959d: 31 runs across 2 machines, 0 validated — model revision and runtime pinned, launch accepted — and 31 candidate, their label for useful evidence without a reproducible-launch promise.

Elsewhere in the registry

Registry record as of 2026-08-13. Identity, licence and parameter count are properties of a release and do not move; tier can, and the verify link above re-checks it against the live registry rather than this page.