BenchAtlas

Rankings / Z.ai

GLM 5.2

open weights

released 2026-06-13

BenchAtlas Index

as of 2026-07-14
76.5
max effort
rank #30
9 families · 5 categories · high
72.7
max effort
rank #52
7 families · 3 categories · medium
63.8
base configuration
rank #108
8 families · 4 categories · medium

Benchmark evidence

67 results
agentic coding
Release 2026-06-25
tasks_counted=3 · livebench_version=2026-06-25
51.9 %independentLiveBench
τ²-Benchmax effort99.1 %independentArtificial Analysis
τ²-Bench
subset=banking
max effort26.8 %independentArtificial Analysis
τ²-Bench
subset=banking
no reasoning15.7 %independentArtificial Analysis
coding
Release 2026-06-25
tasks_counted=2 · livebench_version=2026-06-25
79.7 %independentLiveBench
SciCodemax effort50.5 %independentArtificial Analysis
SciCodeno reasoning36.1 %independentArtificial Analysis
data analysis
Release 2026-06-25
tasks_counted=3 · livebench_version=2026-06-25
73.7 %independentLiveBench
external indices
AA Coding Indexmax effort68.8 pointsindependentArtificial Analysis
AA Coding Indexno reasoning46.5 pointsindependentArtificial Analysis
Intelligence Index v4.1max effort51.1 pointsindependentArtificial Analysis
Intelligence Index v4.1no reasoning34.1 pointsindependentArtificial Analysis
ECI151.6 pointsindependentEpoch AI Benchmarking Hub
ECIhigh effort151.6 pointsindependentEpoch AI Benchmarking Hub
ECImax effort151.6 pointsindependentEpoch AI Benchmarking Hub
Vals Indexmax effort65.0 pointsindependentVals AI
factuality
SimpleQA Verifiedmax effort38.1 %independentEpoch AI Benchmarking Hub
2026-06-19
human preference
Coding (style control)older version
arena=text · category=coding · style_control=true
max effort1510.9
1500.8–1520.9
communityLMArena Leaderboard Dataset
Chinese (style control)older version
arena=text · category=chinese · style_control=true
max effort1503.9
1481.5–1526.4
communityLMArena Leaderboard Dataset
Industry Software And It Services (style control)older version
arena=text · category=industry_software_and_it_services · style_control=true
max effort1499.4
1490.8–1508.1
communityLMArena Leaderboard Dataset
Industry Mathematical (style control)older version
arena=text · category=industry_mathematical · style_control=true
max effort1496.7
1475.3–1518.1
communityLMArena Leaderboard Dataset
Hard Prompts English (style control)older version
arena=text · category=hard_prompts_english · style_control=true
max effort1491.5
1482.1–1500.8
communityLMArena Leaderboard Dataset
Industry Life And Physical And Social Science (style control)older version
arena=text · category=industry_life_and_physical_and_social_science · style_control=true
max effort1488.3
1475.9–1500.6
communityLMArena Leaderboard Dataset
Hard Prompts (style control)older version
arena=text · category=hard_prompts · style_control=true
max effort1487.1
1479.8–1494.4
communityLMArena Leaderboard Dataset
French (style control)older version
arena=text · category=french · style_control=true
max effort1486.7
1460.2–1513.2
communityLMArena Leaderboard Dataset
Expert (style control)older version
arena=text · category=expert · style_control=true
max effort1483.1
1467.3–1498.9
communityLMArena Leaderboard Dataset
Industry Medicine And Healthcare (style control)older version
arena=text · category=industry_medicine_and_healthcare · style_control=true
max effort1480.0
1461.4–1498.7
communityLMArena Leaderboard Dataset
Math (style control)older version
arena=text · category=math · style_control=true
max effort1477.3
1455.4–1499.2
communityLMArena Leaderboard Dataset
Longer Query (style control)older version
arena=text · category=longer_query · style_control=true
max effort1474.7
1466.2–1483.2
communityLMArena Leaderboard Dataset
English (style control)older version
arena=text · category=english · style_control=true
max effort1474.6
1466.7–1482.6
communityLMArena Leaderboard Dataset
Industry Legal And Government (style control)older version
arena=text · category=industry_legal_and_government · style_control=true
max effort1474.4
1456.8–1492.0
communityLMArena Leaderboard Dataset
Exclude Ties (style control)older version
arena=text · category=exclude_ties · style_control=true
max effort1470.7
1462.4–1479.0
communityLMArena Leaderboard Dataset
Multi Turn (style control)older version
arena=text · category=multi_turn · style_control=true
max effort1465.6
1452.8–1478.5
communityLMArena Leaderboard Dataset
Russian (style control)older version
arena=text · category=russian · style_control=true
max effort1465.3
1449.4–1481.1
communityLMArena Leaderboard Dataset
Overall (style control)
arena=text · category=overall · style_control=true
max effort1464.3
1458.0–1470.6
communityLMArena Leaderboard Dataset
Instruction Following (style control)older version
arena=text · category=instruction_following · style_control=true
max effort1460.8
1451.7–1469.9
communityLMArena Leaderboard Dataset
Industry Business And Management And Financial Operations (style control)older version
arena=text · category=industry_business_and_management_and_financial_operations · style_control=true
max effort1455.5
1444.0–1466.9
communityLMArena Leaderboard Dataset
Non English (style control)older version
arena=text · category=non_english · style_control=true
max effort1451.6
1443.9–1459.3
communityLMArena Leaderboard Dataset
Industry Writing And Literature And Language (style control)older version
arena=text · category=industry_writing_and_literature_and_language · style_control=true
max effort1451.1
1440.6–1461.5
communityLMArena Leaderboard Dataset
Creative Writing (style control)older version
arena=text · category=creative_writing · style_control=true
max effort1442.9
1430.5–1455.4
communityLMArena Leaderboard Dataset
Industry Entertainment And Sports And Media (style control)older version
arena=text · category=industry_entertainment_and_sports_and_media · style_control=true
max effort1434.1
1423.1–1445.0
communityLMArena Leaderboard Dataset
knowledge science
GPQA Diamondmax effort91.9 %independentEpoch AI Benchmarking Hub
2026-06-24
GPQA Diamond
implementation=artificial-analysis
max effort89.5 %independentArtificial Analysis
GPQA Diamond
implementation=vals-ai
max effort85.6 %independentVals AI
GPQA Diamond
implementation=artificial-analysis
no reasoning68.6 %independentArtificial Analysis
Humanity's Last Exam
implementation=artificial-analysis
max effort40.1 %independentArtificial Analysis
Humanity's Last Exam
implementation=artificial-analysis
no reasoning8.4 %independentArtificial Analysis
MMLU-Pro
implementation=vals-ai
max effort86.7 %independentVals AI
language
Release 2026-06-25
tasks_counted=3 · livebench_version=2026-06-25
76.2 %independentLiveBench
long context instruction
AA-LCRmax effort71.3 %independentArtificial Analysis
AA-LCRno reasoning37.3 %independentArtificial Analysis
IFBenchmax effort73.3 %independentArtificial Analysis
Release 2026-06-25
tasks_counted=4 · livebench_version=2026-06-25
62.3 %independentLiveBench
professional
CorpFinmax effort66.1 %independentVals AI
LegalBenchmax effort84.1 %independentVals AI
TaxEvalmax effort73.3 %independentVals AI
reasoning math
ARC-AGI-1older version
split=public_eval · model_type=CoT
80.4 %independentARC Prize Leaderboard
ARC-AGI-1older version
split=semi_private · model_type=CoT
77.0 %independentARC Prize Leaderboard
ARC-AGI-2
split=semi_private · model_type=CoT
22.8 %independentARC Prize Leaderboard
ARC-AGI-2
split=public_eval · model_type=CoT
20.8 %independentARC Prize Leaderboard
ARC-AGI-2
split=semi_private · model_type=CoT
0.3 usd_per_taskindependentARC Prize Leaderboard
ARC-AGI-2
split=public_eval · model_type=CoT
0.2 usd_per_taskindependentARC Prize Leaderboard
ARC-AGI-1older version
split=semi_private · model_type=CoT
0.2 usd_per_taskindependentARC Prize Leaderboard
ARC-AGI-1older version
split=public_eval · model_type=CoT
0.2 usd_per_taskindependentARC Prize Leaderboard
Release 2026-06-25
tasks_counted=4 · livebench_version=2026-06-25
89.8 %independentLiveBench
Release 2026-06-25
tasks_counted=4 · livebench_version=2026-06-25
78.6 %independentLiveBench
OTIS Mock AIME 2024–2025max effort86.4 %independentEpoch AI Benchmarking Hub
2026-06-25

Agent + model results

systems, not bare-model scores
agent + model Epoch Inspect harness + GLM 5.2SWE-bench Verified78.7 %independentEpoch AI Benchmarking Hub
agent + model Artificial Analysis harness + GLM 5.2Terminal-Bench 2.177.9 %independentArtificial Analysis
agent + model Artificial Analysis harness + GLM 5.2Terminal-Bench 2.151.7 %independentArtificial Analysis
agent + model Artificial Analysis harness + GLM 5.2Terminal-Bench Hard50.8 %independentArtificial Analysis

These scores measure the whole agent system (scaffold, tools, budgets) — they are never merged into the bare model’s numbers.

Related variants

GLM 5GLM 5.1GLM 4.7GLM 4.6GLM 4.5GLM 4 6vGLM 4 5vGLM 4 0520