Compare · ModelsLive · 2 picked · head to head
Claude Opus 4.6 vs GLM 5.2
Side by side · benchmarks, pricing, and signals you can act on.
Winner summary
Claude Opus 4.6 wins on 14/20 benchmarks
Claude Opus 4.6 wins 14 of 20 shared benchmarks. Leads in reasoning · general · coding.
Category leads
reasoning·Claude Opus 4.6arena·GLM 5.2knowledge·GLM 5.2general·Claude Opus 4.6coding·Claude Opus 4.6math·Claude Opus 4.6
Hype vs Reality
Attention vs performance
Claude Opus 4.6
#136 by perf·#9 by attention
GLM 5.2
#67 by perf·#3 by attention
Best value
GLM 5.2
7.0x better value than Claude Opus 4.6
Claude Opus 4.6
3.2 pts/$
$15.00/M
GLM 5.2
22.6 pts/$
$2.50/M
Vendor risk
Who is behind the model
Anthropic
$965.0B·Tier 1
z-ai
private · undisclosed
Head to head
20 benchmarks · 2 models
Claude Opus 4.6GLM 5.2
ARC-AGI
Claude Opus 4.6 leads by +17.0
ARC-AGI · the original Abstraction and Reasoning Corpus, testing whether AI can solve novel visual pattern recognition tasks without memorization.
Claude Opus 4.6
94.0
GLM 5.2
77.0
ARC-AGI-2
Claude Opus 4.6 leads by +46.4
ARC-AGI-2 · the second iteration of the Abstraction and Reasoning Corpus, testing novel pattern recognition and abstract reasoning without prior training data.
Claude Opus 4.6
69.2
GLM 5.2
22.8
Chatbot Arena Elo · Coding
GLM 5.2 leads by +56.2
Claude Opus 4.6
1537.0
GLM 5.2
1593.3
Chatbot Arena Elo · Overall
Claude Opus 4.6 leads by +26.3
Claude Opus 4.6
1497.3
GLM 5.2
1471.0
Chess Puzzles
GLM 5.2 leads by +4.2
Chess Puzzles · tests strategic and tactical reasoning by having models solve chess puzzle positions, evaluating lookahead and pattern recognition abilities.
Claude Opus 4.6
12.7
GLM 5.2
16.9
Dtbench
GLM 5.2 leads by +4.0
Claude Opus 4.6
85.3
GLM 5.2
89.3
Ebr Bench
Claude Opus 4.6 leads by +3.2
Claude Opus 4.6
12.7
GLM 5.2
9.5
Frontiercode
Claude Opus 4.6 leads by +2.1
Claude Opus 4.6
26.6
GLM 5.2
24.5
FrontierMath-Tier-4-v2-Private
GLM 5.2 leads by +2.4
Claude Opus 4.6
26.8
GLM 5.2
29.3
FrontierMath-Tiers-1-3-v2-Private
Claude Opus 4.6 leads by +6.7
Claude Opus 4.6
66.0
GLM 5.2
59.2
GPQA diamond
GLM 5.2 leads by +1.8
Graduate-Level Google-Proof QA (Diamond set) · expert-crafted questions in physics, biology, and chemistry that are difficult even for domain PhDs.
Claude Opus 4.6
87.4
GLM 5.2
89.1
Lmca
Claude Opus 4.6 leads by +11.7
Claude Opus 4.6
65.6
GLM 5.2
53.9
Mystery Game Puzzles
Claude Opus 4.6 leads by +6.6
Claude Opus 4.6
17.4
GLM 5.2
10.8
OTIS Mock AIME 2024-2025
Claude Opus 4.6 leads by +8.1
OTIS Mock AIME 2024-2025 · simulated American Invitational Mathematics Examination problems testing advanced problem-solving skills.
Claude Opus 4.6
94.4
GLM 5.2
86.4
PostTrainBench
GLM 5.2 leads by +6.9
Claude Opus 4.6
24.8
GLM 5.2
31.7
Proofbench
Claude Opus 4.6 leads by +15.0
Claude Opus 4.6
50.0
GLM 5.2
35.0
SimpleBench
Claude Opus 4.6 leads by +10.6
SimpleBench · tests fundamental reasoning capabilities with straightforward problems designed to expose gaps in basic logical and spatial thinking.
Claude Opus 4.6
61.1
GLM 5.2
50.6
SimpleQA Verified
Claude Opus 4.6 leads by +12.8
SimpleQA Verified · short factual questions with verified answers, measuring factual accuracy and the tendency to hallucinate or provide incorrect information.
Claude Opus 4.6
47.0
GLM 5.2
34.2
SWE-Bench verified
Claude Opus 4.6 leads by +0.0
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. Claude Mythos Preview leads at 93.9%, crossing 90% for the first time in 2026. Opus 4.6 scores 80.8%. The benchmark remains the most-cited evaluation for code-generation capability.
Claude Opus 4.6
78.7
GLM 5.2
78.7
WeirdML
Claude Opus 4.6 leads by +7.8
WeirdML · tests models on unusual and adversarial machine learning tasks that require creative problem-solving beyond standard patterns.
Claude Opus 4.6
78.0
GLM 5.2
70.1
Full benchmark table
| Benchmark | Claude Opus 4.6 | GLM 5.2 |
|---|---|---|
ARC-AGI ARC-AGI · the original Abstraction and Reasoning Corpus, testing whether AI can solve novel visual pattern recognition tasks without memorization. | 94.0 | 77.0 |
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. | 69.2 | 22.8 |
Chatbot Arena Elo · Coding | 1537.0 | 1593.3 |
Chatbot Arena Elo · Overall | 1497.3 | 1471.0 |
Chess Puzzles Chess Puzzles · tests strategic and tactical reasoning by having models solve chess puzzle positions, evaluating lookahead and pattern recognition abilities. | 12.7 | 16.9 |
Dtbench | 85.3 | 89.3 |
Ebr Bench | 12.7 | 9.5 |
Frontiercode | 26.6 | 24.5 |
FrontierMath-Tier-4-v2-Private | 26.8 | 29.3 |
FrontierMath-Tiers-1-3-v2-Private | 66.0 | 59.2 |
GPQA diamond Graduate-Level Google-Proof QA (Diamond set) · expert-crafted questions in physics, biology, and chemistry that are difficult even for domain PhDs. | 87.4 | 89.1 |
Lmca | 65.6 | 53.9 |
Mystery Game Puzzles | 17.4 | 10.8 |
OTIS Mock AIME 2024-2025 OTIS Mock AIME 2024-2025 · simulated American Invitational Mathematics Examination problems testing advanced problem-solving skills. | 94.4 | 86.4 |
PostTrainBench | 24.8 | 31.7 |
Proofbench | 50.0 | 35.0 |
SimpleBench SimpleBench · tests fundamental reasoning capabilities with straightforward problems designed to expose gaps in basic logical and spatial thinking. | 61.1 | 50.6 |
SimpleQA Verified SimpleQA Verified · short factual questions with verified answers, measuring factual accuracy and the tendency to hallucinate or provide incorrect information. | 47.0 | 34.2 |
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. Claude Mythos Preview leads at 93.9%, crossing 90% for the first time in 2026. Opus 4.6 scores 80.8%. The benchmark remains the most-cited evaluation for code-generation capability. | 78.7 | 78.7 |
WeirdML WeirdML · tests models on unusual and adversarial machine learning tasks that require creative problem-solving beyond standard patterns. | 78.0 | 70.1 |
Pricing · per 1M tokens · projected $/mo at 10M tokens
| Model | Input | Output | Context | Projected $/mo |
|---|---|---|---|---|
| $5.00 | $25.00 | 1.0M tokens (~500 books) | $100.00 | |
| $1.00 | $4.00 | 1.0M tokens (~524 books) | $17.50 |