AI Model Release Tracker - Timeline of Major AI Models from 2022-2026

Kimi K3 vs GLM-5.2

Moonshot AI
Kimi K3
Z.ai
GLM-5.2
Overview
CompanyMoonshot AIZ.ai
Release dateJul 16 2026Jun 16 2026
AccessOpen WeightOpen Weight
Specifications
Parameters
2.8T
744B
Context window
1M
1M
Benchmarks
Nonsense detection
BullshitBench v2
31%
Agentic coding
SWE-Bench Pro
62.1%
Agentic coding
CursorBench v3.1
54.6%
Agentic coding
DeepSWE 1.0
67.5%
Next.js coding
Next.js Evals
88%
Agentic terminal coding
Terminal-Bench 2.1
88.3%Best
81%
Multi-step tool use
MCP Atlas
84.2%
Professional tool use
JobBench
52.9%
Personal tool use
Toolathlon-Verified
73.2%
Web browsing
BrowseComp
91.2%
Multidisciplinary reasoning
Humanity's Last Exam · no tools
43.5%Best
40.5%
Multidisciplinary reasoning
Humanity's Last Exam · with tools
56%Best
54.7%
Science
GPQA Diamond
93.5%Best
91.2%
Knowledge work
GDPval-AA v2
1668Best
1514
Chart reasoning
CharXiv Reasoning
84.8%
Multimodal reasoning
MMMU-Pro
81.6%
Community preference (code)
Arena Elo (Code)
1580
Timeline
Release gapGLM-5.2 shipped 30 days before Kimi K3

Which is better: Kimi K3 or GLM-5.2?

Kimi K3 leads GLM-5.2 on 5 of the 5 benchmarks they both report (Terminal-Bench 2.1, Humanity's Last Exam, GPQA Diamond, GDPval-AA v2). GLM-5.2 shipped 30 days before Kimi K3, so benchmark comparisons should account for the intervening progress.

Kimi K3 has 2.8T parameters, while GLM-5.2 has 744B. Context windows are 1M (Kimi K3) vs 1M (GLM-5.2).

On Terminal-Bench 2.1, Kimi K3 leads at 88.3% vs GLM-5.2 at 81%. On Humanity's Last Exam · no tools, Kimi K3 leads at 43.5% vs GLM-5.2 at 40.5%. On Humanity's Last Exam · with tools, Kimi K3 leads at 56% vs GLM-5.2 at 54.7%. On GPQA Diamond, Kimi K3 leads at 93.5% vs GLM-5.2 at 91.2%. On GDPval-AA v2, Kimi K3 leads at 1668 vs GLM-5.2 at 1514.

Frequently asked questions

Kimi K3 was released by Moonshot AI on Jul 16 2026.

GLM-5.2 was released by Z.ai on Jun 16 2026.

Kimi K3 leads on Terminal-Bench 2.1 — Kimi K3 88.3% vs GLM-5.2 81%.

Kimi K3 leads on Humanity's Last Exam · no tools — Kimi K3 43.5% vs GLM-5.2 40.5%.

Kimi K3 has a 1M context window; GLM-5.2 has 1M.

Other comparisons