Compare · ModelsLive · 2 picked · head to head
GLM 5.2 vs GPT-5
Side by side · benchmarks, pricing, and signals you can act on.
Winner summary
GLM 5.2 wins on 12/17 benchmarks
GLM 5.2 wins 12 of 17 shared benchmarks. Leads in agentic · reasoning · general.
Category leads
agentic·GLM 5.2reasoning·GLM 5.2knowledge·GPT-5general·GLM 5.2math·GLM 5.2coding·GLM 5.2
Hype vs Reality
Attention vs performance
GLM 5.2
#75 by perf·#3 by attention
GPT-5
#116 by perf·#9 by attention
Best value
GLM 5.2
1.8x better value than GPT-5
GLM 5.2
16.5 pts/$
$3.40/M
GPT-5
9.3 pts/$
$5.63/M
Vendor risk
Who is behind the model
z-ai
private · undisclosed
OpenAI
$840.0B·Tier 1
Head to head
17 benchmarks · 2 models
GLM 5.2GPT-5
APEX-Agents
GLM 5.2 leads by +26.9
APEX-Agents · evaluates AI agents on complex, multi-step tasks requiring planning, tool use, and autonomous decision-making in realistic environments.
GLM 5.2
45.2
GPT-5
18.3
ARC-AGI
GLM 5.2 leads by +11.3
ARC-AGI · the original Abstraction and Reasoning Corpus, testing whether AI can solve novel visual pattern recognition tasks without memorization.
GLM 5.2
77.0
GPT-5
65.7
ARC-AGI-2
GLM 5.2 leads by +12.9
ARC-AGI-2 · the second iteration of the Abstraction and Reasoning Corpus, testing novel pattern recognition and abstract reasoning without prior training data.
GLM 5.2
22.8
GPT-5
9.9
Chess Puzzles
GPT-5 leads by +19.1
Chess Puzzles · tests strategic and tactical reasoning by having models solve chess puzzle positions, evaluating lookahead and pattern recognition abilities.
GLM 5.2
17.9
GPT-5
37.0
Dtbench
GLM 5.2 leads by +4.9
GLM 5.2
89.3
GPT-5
84.5
Ebr Bench
GPT-5 leads by +3.2
GLM 5.2
9.5
GPT-5
12.7
FrontierMath-Tier-4-v2-Private
GLM 5.2 leads by +7.3
GLM 5.2
29.3
GPT-5
21.9
FrontierMath-Tiers-1-3-v2-Private
GLM 5.2 leads by +4.0
GLM 5.2
62.3
GPT-5
58.4
GPQA diamond
GLM 5.2 leads by +7.9
Graduate-Level Google-Proof QA (Diamond set) · expert-crafted questions in physics, biology, and chemistry that are difficult even for domain PhDs.
GLM 5.2
92.9
GPT-5
85.0
Lmca
GLM 5.2 leads by +6.9
GLM 5.2
53.9
GPT-5
47.0
Mystery Game Puzzles
GPT-5 leads by +4.4
GLM 5.2
10.8
GPT-5
15.2
OTIS Mock AIME 2024-2025
GPT-5 leads by +5.0
OTIS Mock AIME 2024-2025 · simulated American Invitational Mathematics Examination problems testing advanced problem-solving skills.
GLM 5.2
86.4
GPT-5
91.4
Proofbench
GLM 5.2 leads by +17.0
GLM 5.2
35.0
GPT-5
18.0
SimpleBench
GLM 5.2 leads by +2.5
SimpleBench · tests fundamental reasoning capabilities with straightforward problems designed to expose gaps in basic logical and spatial thinking.
GLM 5.2
50.6
GPT-5
48.0
SimpleQA Verified
GPT-5 leads by +16.4
SimpleQA Verified · short factual questions with verified answers, measuring factual accuracy and the tendency to hallucinate or provide incorrect information.
GLM 5.2
34.2
GPT-5
50.6
SWE-Bench verified
GLM 5.2 leads by +5.2
SWE-bench Verified · 500 human-validated tasks from 12 real Python repositories (Django, Flask, scikit-learn, sympy, and others). Each task requires the model to produce a git patch that resolves a real GitHub issue and passes the test suite. The verified subset eliminates ambiguous tasks from the original SWE-bench. The benchmark remains the most-cited evaluation for code-generation capability.
GLM 5.2
78.7
GPT-5
73.5
WeirdML
GLM 5.2 leads by +9.4
WeirdML · tests models on unusual and adversarial machine learning tasks that require creative problem-solving beyond standard patterns.
GLM 5.2
70.1
GPT-5
60.7
Full benchmark table
| Benchmark | GLM 5.2 | GPT-5 |
|---|---|---|
APEX-Agents APEX-Agents · evaluates AI agents on complex, multi-step tasks requiring planning, tool use, and autonomous decision-making in realistic environments. | 45.2 | 18.3 |
ARC-AGI ARC-AGI · the original Abstraction and Reasoning Corpus, testing whether AI can solve novel visual pattern recognition tasks without memorization. | 77.0 | 65.7 |
ARC-AGI-2 ARC-AGI-2 · the second iteration of the Abstraction and Reasoning Corpus, testing novel pattern recognition and abstract reasoning without prior training data. | 22.8 | 9.9 |
Chess Puzzles Chess Puzzles · tests strategic and tactical reasoning by having models solve chess puzzle positions, evaluating lookahead and pattern recognition abilities. | 17.9 | 37.0 |
Dtbench | 89.3 | 84.5 |
Ebr Bench | 9.5 | 12.7 |
FrontierMath-Tier-4-v2-Private | 29.3 | 21.9 |
FrontierMath-Tiers-1-3-v2-Private | 62.3 | 58.4 |
GPQA diamond Graduate-Level Google-Proof QA (Diamond set) · expert-crafted questions in physics, biology, and chemistry that are difficult even for domain PhDs. | 92.9 | 85.0 |
Lmca | 53.9 | 47.0 |
Mystery Game Puzzles | 10.8 | 15.2 |
OTIS Mock AIME 2024-2025 OTIS Mock AIME 2024-2025 · simulated American Invitational Mathematics Examination problems testing advanced problem-solving skills. | 86.4 | 91.4 |
Proofbench | 35.0 | 18.0 |
SimpleBench SimpleBench · tests fundamental reasoning capabilities with straightforward problems designed to expose gaps in basic logical and spatial thinking. | 50.6 | 48.0 |
SimpleQA Verified SimpleQA Verified · short factual questions with verified answers, measuring factual accuracy and the tendency to hallucinate or provide incorrect information. | 34.2 | 50.6 |
SWE-Bench verified SWE-bench Verified · 500 human-validated tasks from 12 real Python repositories (Django, Flask, scikit-learn, sympy, and others). Each task requires the model to produce a git patch that resolves a real GitHub issue and passes the test suite. The verified subset eliminates ambiguous tasks from the original SWE-bench. The benchmark remains the most-cited evaluation for code-generation capability. | 78.7 | 73.5 |
WeirdML WeirdML · tests models on unusual and adversarial machine learning tasks that require creative problem-solving beyond standard patterns. | 70.1 | 60.7 |
Pricing · per 1M tokens · projected $/mo at 10M tokens