RankingEffort benchmarks

GLM-5.2 · Low

Compare models

67 measured results · 7 effort settings · 46 benchmark/harness combinations

Model evidence and sources →

Change model

Model

Current configurationSources

Compare with

Choose a second model and its effort to see both profiles and benchmark differences below.

Compare withModel and configuration

Capability profile

One configuration, seven capabilities

Comparison score · 0–100

AGLM-5.2 · Low1 of 7 capabilities supported

Gaps are unknown, not zero. Hollow points are preliminary.

Agentic
Unknown
Hard reasoning
A28.8Supported
Coding
Unknown
Human pref
Unknown
Knowledge
Unknown
Multimodal
Unknown
Long context
Unknown

Each profile belongs to the named configuration. Capability scores use a shared panel for all axes, distinct from the overall ranking panel; they are not accuracy percentages or direct head-to-head wins.

Capability score evidence

GLM-5.2 · Low: 18.7 overall comparison score. Meets the overall evidence requirements.

Hard reasoning only. The available task mix affects this estimate; missing areas remain unknown.

Which sources carry the weight?

Epoch AI · 100.0% direct comparison weight · independent of the model provider

Which benchmark families carry the weight?

GPQA · 29.2% direct comparison weight
Aime · 28.9% direct comparison weight
epoch-game-puzzles · 26.5% direct comparison weight
FrontierMath · 15.5% direct comparison weight

These shares describe direct comparison weight, not fractions of the final score or confidence. Source balancing does not give every family equal weight or remove correlations between overlapping sources.

How sensitive is this score?

The audit removes one source at a time and refits against the same overall reference panel. These published results use the matching dataset audited 2026-10-23.

Both ranks use the same surviving configurations for each removal. The cohort starts with 444 configurations supported in the original audit, including separate effort settings. These are diagnostic positions, not the default leaderboard ranks. Eligibility is not rechecked after removal.

What changes when a source is removed?
Removed source	Rank before → after	Rank movement	Cohort remaining / lost support	Score after (change)
Artificial Analysis	123 → 120	Up 3	203 remaining / 241 lost	17.8 (-0.9 points)
ARC Prize	144 → 144	No movement	444 remaining / 0 lost	22.7 (+4.0 points)
Datacurve	144 → 145	Down 1	444 remaining / 0 lost	15.3 (-3.4 points)
LiveCodeBench	144 → 144	No movement	444 remaining / 0 lost	18.7 (-0.0 points)
Scale	144 → 144	No movement	444 remaining / 0 lost	20.6 (+1.9 points)
Terminal-Bench	144 → 144	No movement	444 remaining / 0 lost	18.7 (+0.0 points)
Cognition	144 → 136	Up 8	444 remaining / 0 lost	19.0 (+0.3 points)
LiveBench	143 → 145	Down 2	443 remaining / 1 lost	18.7 (+0.0 points)
SWE-rebench	144 → 143	Up 1	444 remaining / 0 lost	19.0 (+0.3 points)
Epoch AI	Comparison support lost; no rank		431 remaining / 13 lost	Unavailable
Vals AI	144 → 142	Up 2	443 remaining / 1 lost	17.9 (-0.8 points)
Provider reports · Anthropic	144 → 144	No movement	444 remaining / 0 lost	18.7 (-0.0 points)
Provider reports · Moonshot AI	144 → 144	No movement	444 remaining / 0 lost	18.3 (-0.4 points)
Provider reports · Meta	144 → 144	No movement	444 remaining / 0 lost	18.7 (-0.0 points)
Provider reports · Mistral	144 → 144	No movement	444 remaining / 0 lost	18.7 (+0.0 points)
Provider reports · OpenAI	144 → 144	No movement	444 remaining / 0 lost	18.8 (+0.1 points)
All provider reports	144 → 144	No movement	444 remaining / 0 lost	18.4 (-0.3 points)

Lost support means a configuration no longer has a comparison path to the complete reference panel. Its rank is unavailable. A surviving path alone does not establish leaderboard eligibility.

Overlapping sources may remain correlated even with source balancing. Leave-one-source-out movement is a sensitivity diagnostic, not a spread of source-specific ranks or a confidence measure. Scoring methodology · Dated configuration audit

Every benchmark family

Removed evidence	Recomputed score	Change
aime	26.1	+7.4
gpqa	15.2	-3.5
deepswe	15.3	-3.4
arc-agi	22.0	+3.3
epoch-game-puzzles	15.9	-2.8
frontiermath	17.4	-1.3
mmmu-pro	19.8	+1.0
hle	19.6	+0.9
simpleqa	19.3	+0.6
tau-bench	18.2	-0.5
mmlu-pro	19.2	+0.5
livebench-instructions	19.0	+0.3
frontiercode	19.0	+0.3
multi-swe-bench	18.4	-0.3
swe-rebench	19.0	+0.3
omniscience	18.5	-0.2
swe-atlas-qna	18.5	-0.2
livebench-coding	18.5	-0.2
livebench-language	18.9	+0.2
harvey	18.5	-0.2
arc-agi-3	18.5	-0.2
ifbench	18.9	+0.2
apex-agents	18.9	+0.2
gdp-pdf	18.8	+0.1
gdpval	18.6	-0.1
terminal-bench-science	18.8	+0.1
enigma-eval	18.6	-0.1
livebench-reasoning	18.6	-0.1
vibe-code	18.6	-0.1
itbench	18.6	-0.1
analyst-agent	18.8	+0.1
gmmlu	18.8	+0.1
vals-legal-research	18.6	-0.1
briefcase	18.6	-0.1
automationbench	18.6	-0.1
vals-finance-agent	18.6	-0.1
swe-atlas-test-writing	18.6	-0.1
livebench-math	18.8	+0.1
critpt	18.7	-0.0
vals-tax-agent	18.7	-0.0
vals-code-migration	18.7	-0.0
terminal-bench	18.7	-0.0
math500	18.7	+0.0
aa-lcr	18.7	+0.0
vals-excel-modeling	18.7	-0.0
enterprise-ops	18.7	+0.0
livebench-data	18.7	+0.0
livecodebench	18.7	+0.0
swe-atlas-refactoring	18.7	-0.0
swe-bench-pro	18.7	+0.0

What evidence is missing from the fit?

Across all collected settings for this model: 127 observations, 59 contributing, 0 matched without graph weight and 68 excluded. Counts do not establish rank eligibility.

5 · Benchmark family or source protocol has not been reviewed
1 · Original SciCode grading defects are under review; retained as supporting evidence, excluded from aggregate and capability scores. Corrected SciCode-Verified requires a separate protocol review. Sources: https://arxiv.org/abs/2608.04975 and https://artificialanalysis.ai/evaluations/scicode
2 · The evaluator reports no named effort control. This does not identify a non-reasoning configuration.
3 · Effort is not identified
3 · Evaluator has not been reviewed for independent admission
14 · No reviewed effort for this catalog observation; any separately collected effort measurement is counted separately
17 · Supporting evidence outside the reviewed capability core
7 · Kimi K3 Evaluation Details cites Artificial Analysis; retain as published context, not a new Moonshot comparison.
1 · Kimi K3 Evaluation Details cites GLM release blog or Vals AI, per model; retain as published context, not a new Moonshot comparison.
1 · Kimi K3 Evaluation Details cites GLM release blog, Artificial Analysis or OpenAI, per model; retain as published context, not a new Moonshot comparison.
1 · Kimi K3 Evaluation Details cites FrontierSWE leaderboard; retain as published context, not a new Moonshot comparison. Historical provider claim; secondary and superseded for current FrontierSWE comparisons by the independent Proximal FrontierSWE v2 mean@5 board at https://www.frontierswe.com/. Version-unspecified claims are not v2 results. Excluded from weighted Overall.
1 · Kimi K3 Evaluation Details cites GLM-5.2 release blog; retain as published context, not a new Moonshot comparison.
1 · Kimi K3 Evaluation Details cites official PostTrainBench results; retain as published context, not a new Moonshot comparison.
1 · Kimi K3 Evaluation Details cites official Agents Last Exam leaderboard; retain as published context, not a new Moonshot comparison.
1 · Kimi K3 Evaluation Details cites Artificial Analysis / APEX-Agents leaderboard; retain as published context, not a new Moonshot comparison.
3 · Kimi K3 Evaluation Details cites Vals AI; retain as published context, not a new Moonshot comparison.
1 · The GLM card attributes this evaluation to Proximal; it is not a new Z.ai comparison. Historical provider claim; secondary and superseded for current FrontierSWE comparisons by the independent Proximal FrontierSWE v2 mean@5 board at https://www.frontierswe.com/. Version-unspecified claims are not v2 results. Excluded from weighted Overall.
4 · Narrow specialist biological, medical or security evaluation; preserved as raw evidence outside general capability scoring.
1 · The GLM card attributes this evaluation to Artificial Analysis; it is not a new Z.ai comparison.

Inspect every observation and exclusion →

Benchmark results

These are published benchmark scores. Supported effort estimates also appear on the Capability leaderboard. Results from different harnesses occupy different rows. Missing cells stay unknown; no scores are copied between effort levels.

Benchmark / source	Minimal	Low	Medium	High	Max	Non-reasoning	Unspecified
AA-Briefcase EloArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details AA-Briefcase v1.1 is an agentic knowledge work benchmark developed by Artificial Analysis. AA-Briefcase Elo is a combined metric that aggregates rubric pass rate, analytical quality Elo and presentation Elo · Higher is better · Evaluation results measured independently by Artificial Analysis Unit: elo. Source collected 2026-10-18.	Not reported	Not reported	Not reported	Not reported	1,228.86 Elo Reported settings GLM-5.2 (Max); Model source: https://artificialanalysis.ai/models/glm-5-2	Not reported	Not reported
APEX-Agents-AAArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-10-18.	Not reported	Not reported	Not reported	Not reported	33.7% Reported settings GLM-5.2 (Max); Model source: https://artificialanalysis.ai/models/glm-5-2	Not reported	Not reported
AA-LCR v1.1Artificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-10-18.	Not reported	Not reported	Not reported	Not reported	78.33% Reported settings GLM-5.2 (Max); Model source: https://artificialanalysis.ai/models/glm-5-2	42.33% Reported settings GLM-5.2 (Non-reasoning); Model source: https://artificialanalysis.ai/models/glm-5-2-non-reasoning	Not reported
AutomationBench-AAArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Share of task objectives completed with no guardrail violations · Higher is better · Benchmark developed by Zapier · Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-10-18.	Not reported	Not reported	Not reported	Not reported	28.4% Reported settings GLM-5.2 (Max); Model source: https://artificialanalysis.ai/models/glm-5-2	Not reported	Not reported
CritPtArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Benchmark developed by Argonne and UIUC, with contributions from 60+ researchers globally. CritPt is under review by its authors following external feedback · Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-10-18.	Not reported	Not reported	Not reported	Not reported	20.86% Reported settings GLM-5.2 (Max); Model source: https://artificialanalysis.ai/models/glm-5-2	3.14% Reported settings GLM-5.2 (Non-reasoning); Model source: https://artificialanalysis.ai/models/glm-5-2-non-reasoning	Not reported
EnterpriseOps-Gym-AAArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Overall task success rate across all eight enterprise domains (oracle tool mode) · Benchmark developed by ServiceNow Research · Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-10-18.	Not reported	Not reported	Not reported	Not reported	42.73% Reported settings GLM-5.2 (Max); Model source: https://artificialanalysis.ai/models/glm-5-2	Not reported	Not reported
GDP.pdf: All-pass RateArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Share of attempts where every atomic criterion passed · Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-10-18.	Not reported	Not reported	Not reported	Not reported	10.4% Reported settings GLM-5.2 (Max); Model source: https://artificialanalysis.ai/models/glm-5-2	Not reported	Not reported
GDPval-AA v2.1 LeaderboardArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Elo rating for performance on real-world work tasks · Anchored to DeepSeek V4.1 Flash (max) at 1600 · Higher is better · Evaluation results measured independently by Artificial Analysis Unit: elo. Source collected 2026-10-18.	Not reported	Not reported	Not reported	Not reported	1,372.01 Elo Reported settings GLM-5.2 (Max); Model source: https://artificialanalysis.ai/models/glm-5-2	1,247.03 Elo Reported settings GLM-5.2 (Non-reasoning); Model source: https://artificialanalysis.ai/models/glm-5-2-non-reasoning	Not reported
GPQA DiamondArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-10-18.	Not reported	Not reported	Not reported	Not reported	89.49% Reported settings GLM-5.2 (Max); Model source: https://artificialanalysis.ai/models/glm-5-2	68.59% Reported settings GLM-5.2 (Non-reasoning); Model source: https://artificialanalysis.ai/models/glm-5-2-non-reasoning	Not reported
Humanity's Last ExamArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-10-18.	Not reported	Not reported	Not reported	Not reported	41.15% Reported settings GLM-5.2 (Max); Model source: https://artificialanalysis.ai/models/glm-5-2	9.78% Reported settings GLM-5.2 (Non-reasoning); Model source: https://artificialanalysis.ai/models/glm-5-2-non-reasoning	Not reported
IFBenchArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-10-18.	Not reported	Not reported	Not reported	Not reported	73.33% Reported settings GLM-5.2 (Max); Model source: https://artificialanalysis.ai/models/glm-5-2	Not reported	Not reported
ITBench-AAArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Combined public and private Site Reliability Engineering (SRE) tasks · Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-10-18.	Not reported	Not reported	Not reported	Not reported	42.66% Reported settings GLM-5.2 (Max); Model source: https://artificialanalysis.ai/models/glm-5-2	Not reported	Not reported
MLCR-AAArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Share of tasks judged accurate, complete, and concise · Expert + compound (hard) sets · Benchmark developed by Wisedocs · Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-10-18.	Not reported	Not reported	Not reported	Not reported	7.22% Reported settings GLM-5.2 (Max); Model source: https://artificialanalysis.ai/models/glm-5-2	Not reported	Not reported
AA-Omniscience IndexArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details AA-Omniscience Index (higher is better) measures knowledge reliability and hallucination. It rewards correct answers, penalizes hallucinations, and has no penalty for refusing to answer. Scores range from -100 to 100, where 0 means as many correct as incorrect answers, and negative scores mean more incorrect than correct. · Evaluation results measured independently by Artificial Analysis Unit: index. Source collected 2026-10-18.	Not reported	Not reported	Not reported	Not reported	4.43 Reported settings GLM-5.2 (Max); Model source: https://artificialanalysis.ai/models/glm-5-2	-6.57 Reported settings GLM-5.2 (Non-reasoning); Model source: https://artificialanalysis.ai/models/glm-5-2-non-reasoning	Not reported
AA-Omniscience AccuracyArtificial Analysis · 6000 questions; 1 attempt; correct-answer share; GPT-5.6 Luna medium grader Benchmark details AA-Omniscience Accuracy (higher is better) measures the proportion of correctly answered questions out of all questions, regardless of whether the model chooses to answer · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-10-18.	Not reported	Not reported	Not reported	Not reported	24.33% Reported settings GLM-5.2 (Max); Model source: https://artificialanalysis.ai/models/glm-5-2	20.28% Reported settings GLM-5.2 (Non-reasoning); Model source: https://artificialanalysis.ai/models/glm-5-2-non-reasoning	Not reported
SciCodeArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Benchmark curated by scientists across 16 disciplines. SciCode is under review following an independent audit of the dataset · Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-10-18.	Not reported	Not reported	Not reported	Not reported	51.16% Reported settings GLM-5.2 (Max); Model source: https://artificialanalysis.ai/models/glm-5-2	Not reported	Not reported
𝜏²-Bench TelecomArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-10-18.	Not reported	Not reported	Not reported	Not reported	99.12% Reported settings GLM-5.2 (Max); Model source: https://artificialanalysis.ai/models/glm-5-2	Not reported	Not reported
𝜏³-BankingArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Benchmark developed by Sierra Research · Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-10-18.	Not reported	Not reported	Not reported	Not reported	34.64% Reported settings GLM-5.2 (Max); Model source: https://artificialanalysis.ai/models/glm-5-2	16.7% Reported settings GLM-5.2 (Non-reasoning); Model source: https://artificialanalysis.ai/models/glm-5-2-non-reasoning	Not reported
Terminal-Bench HardArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-10-18.	Not reported	Not reported	Not reported	Not reported	50.76% Reported settings GLM-5.2 (Max); Model source: https://artificialanalysis.ai/models/glm-5-2	Not reported	Not reported
Terminal-Bench v2.1Artificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Benchmark developed by the Laude Institute and open-source contributors · Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-10-18.	Not reported	Not reported	Not reported	Not reported	77.9% Reported settings GLM-5.2 (Max); Model source: https://artificialanalysis.ai/models/glm-5-2	51.69% Reported settings GLM-5.2 (Non-reasoning); Model source: https://artificialanalysis.ai/models/glm-5-2-non-reasoning	Not reported
Terminal-Bench v4.0Artificial Analysis · mini-SWE-agent v2.4.6; 66 tasks; pass@1 averaged over 3 repeats; 500 steps; task timeout up to 8 hours Benchmark details Benchmark developed by the Laude Institute, Stanford, and open-source contributors · Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-10-18.	Not reported	Not reported	Not reported	Not reported	1.01% Reported settings GLM-5.2 (Max); Model source: https://artificialanalysis.ai/models/glm-5-2	Not reported	Not reported
DeepSWE v1.1Datacurve · mini-swe-agent Benchmark details Original public board. Configuration and task coverage may differ; inspect source. Unit: percent. Source collected 2026-10-18.	Not reported	Not reported	Not reported	36.3% Reported settings glm-5-2; Reported effort: high; 113 unique tasks; 452 attempts; 4 runs	43.8% Reported settings glm-5-2; Reported effort: max; 113 unique tasks; 450 attempts; 4 runs	Not reported	Not reported
FrontierCode 1.1 · extended-mini-swe-agentCognition · mini-swe-agent Benchmark details FrontierCode 1.1; mini-swe-agent. Open each result for its exact effort, task subset, metric and protocol details. Unit: percent. Source collected 2026-10-18.	Not reported	Not reported	Not reported	Not reported	Not reported	Not reported	40.05% Reported settings FrontierCode 1.1; extended; 150 tasks; mini-swe-agent; source effort none; weighted rubric score, not pass rate; unfair internet runs zeroed by evaluator; flag rate 0 The evaluator reports no named effort control. This does not identify a non-reasoning configuration. Source label needs review
FrontierCode 1.1 · main-mini-swe-agentCognition · mini-swe-agent Benchmark details FrontierCode 1.1; mini-swe-agent. Open each result for its exact effort, task subset, metric and protocol details. Unit: percent. Source collected 2026-10-18.	Not reported	Not reported	Not reported	Not reported	Not reported	Not reported	24.5% Reported settings FrontierCode 1.1; main; 100 tasks; mini-swe-agent; source effort none; weighted rubric score, not pass rate; unfair internet runs zeroed by evaluator; flag rate 0 The evaluator reports no named effort control. This does not identify a non-reasoning configuration. Source label needs review
SWE-rebench · 2026-03 · toolsSWE-rebench · SWE-rebench standardized ReAct tools Benchmark details Average resolved rate across repeated runs, not best-of-N success. Disjoint complete calendar months within each model’s tested range; standardized harness and all-language task set. Unit: percent. Source collected 2026-10-18.	Not reported	Not reported	Not reported	53.48% Reported settings SWE-rebench standardized ReAct tools; swe-rebench-2026-03-tools; GLM-5.2 [high]; checkpoint=glm-5.2	Not reported	Not reported	Not reported

What these comparisons mean

Each source reports its own effort labels. Matching labels do not establish equal compute budgets or identical fallback behavior. Open Reported settings to inspect the exact model name, agent and task coverage. Multiple results at one level remain visible rather than selecting the best. Source-reported None or Non-reasoning labels are not treated as evidence that an API supports those settings.

These measurements use their own dated source collection. The leaderboard fits identified configurations jointly; it contains no pooled model ratings. Explicit thinking budgets are separate configurations. Read the methodology.

Download all effort benchmark results · 8,916 results across 259 models · Collected 2026-10-18

GLM-5.2 · Low

Compare with

One configuration, seven capabilities

Agentic

Hard reasoning

Coding

Human pref

Knowledge

Multimodal

Long context

Capability score evidence

Benchmark results