Compare · ModelsLive · 2 picked · head to head
GLM 5 vs Claude Opus 4.5
Side by side · benchmarks, pricing, and signals you can act on.
Winner summary
Claude Opus 4.5 wins on 13/14 benchmarks
Claude Opus 4.5 wins 13 of 14 shared benchmarks. Leads in reasoning · arena · knowledge.
Category leads
reasoning·Claude Opus 4.5arena·Claude Opus 4.5knowledge·Claude Opus 4.5math·Claude Opus 4.5coding·Claude Opus 4.5
Hype vs Reality
Attention vs performance
GLM 5
#55 by perf·#27 by attention
Claude Opus 4.5
#113 by perf·no signal
Best value
GLM 5
15.1x better value than Claude Opus 4.5
GLM 5
45.7 pts/$
$1.26/M
Claude Opus 4.5
3.0 pts/$
$15.00/M
Vendor risk
Who is behind the model
z-ai
private · undisclosed
Anthropic
$380.0B·Tier 1
Head to head
14 benchmarks · 2 models
GLM 5Claude Opus 4.5
ARC-AGI
Claude Opus 4.5 leads by +35.3
ARC-AGI · the original Abstraction and Reasoning Corpus, testing whether AI can solve novel visual pattern recognition tasks without memorization.
GLM 5
44.7
Claude Opus 4.5
80.0
ARC-AGI-2
Claude Opus 4.5 leads by +32.8
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
4.9
Claude Opus 4.5
37.6
Chatbot Arena Elo · Coding
Claude Opus 4.5 leads by +24.3
GLM 5
1441.0
Claude Opus 4.5
1465.2
Chatbot Arena Elo · Overall
Claude Opus 4.5 leads by +12.2
GLM 5
1455.6
Claude Opus 4.5
1467.7
Chess Puzzles
Claude Opus 4.5 leads by +2.0
Chess Puzzles · tests strategic and tactical reasoning by having models solve chess puzzle positions, evaluating lookahead and pattern recognition abilities.
GLM 5
10.0
Claude Opus 4.5
12.0
FrontierMath-2025-02-28-Private
Claude Opus 4.5 leads by +4.3
FrontierMath (Feb 2025) · original research-level math problems created by mathematicians, testing capabilities at the boundary of current AI mathematical reasoning.
GLM 5
16.4
Claude Opus 4.5
20.7
FrontierMath-Tier-4-2025-07-01-Private
Claude Opus 4.5 leads by +2.1
FrontierMath Tier 4 (Jul 2025) · the most challenging tier of frontier mathematics, containing problems that push the absolute limits of AI mathematical reasoning.
GLM 5
2.1
Claude Opus 4.5
4.2
GPQA diamond
GLM 5 leads by +2.4
Graduate-Level Google-Proof QA (Diamond set) · expert-crafted questions in physics, biology, and chemistry that are difficult even for domain PhDs.
GLM 5
83.8
Claude Opus 4.5
81.4
OTIS Mock AIME 2024-2025
Claude Opus 4.5 leads by +6.1
OTIS Mock AIME 2024-2025 · simulated American Invitational Mathematics Examination problems testing advanced problem-solving skills.
GLM 5
80.0
Claude Opus 4.5
86.1
PostTrainBench
Claude Opus 4.5 leads by +3.4
GLM 5
13.9
Claude Opus 4.5
17.3
SimpleBench
Claude Opus 4.5 leads by +10.6
SimpleBench · tests fundamental reasoning capabilities with straightforward problems designed to expose gaps in basic logical and spatial thinking.
GLM 5
43.8
Claude Opus 4.5
54.4
SWE-Bench verified
Claude Opus 4.5 leads by +4.6
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.
GLM 5
72.1
Claude Opus 4.5
76.7
Terminal Bench
Claude Opus 4.5 leads by +10.7
Terminal-Bench 2.0 · evaluates AI agents on real terminal-based coding tasks · writing scripts, debugging, running tests, and managing projects entirely through command-line interaction. Tests both code quality and terminal fluency. Claude Opus 4.7 scores 69.4%, demonstrating significant agentic terminal competence.
GLM 5
52.4
Claude Opus 4.5
63.1
WeirdML
Claude Opus 4.5 leads by +15.5
WeirdML · tests models on unusual and adversarial machine learning tasks that require creative problem-solving beyond standard patterns.
GLM 5
48.2
Claude Opus 4.5
63.7
Full benchmark table
| Benchmark | GLM 5 | Claude Opus 4.5 |
|---|---|---|
ARC-AGI ARC-AGI · the original Abstraction and Reasoning Corpus, testing whether AI can solve novel visual pattern recognition tasks without memorization. | 44.7 | 80.0 |
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. | 4.9 | 37.6 |
Chatbot Arena Elo · Coding | 1441.0 | 1465.2 |
Chatbot Arena Elo · Overall | 1455.6 | 1467.7 |
Chess Puzzles Chess Puzzles · tests strategic and tactical reasoning by having models solve chess puzzle positions, evaluating lookahead and pattern recognition abilities. | 10.0 | 12.0 |
FrontierMath-2025-02-28-Private FrontierMath (Feb 2025) · original research-level math problems created by mathematicians, testing capabilities at the boundary of current AI mathematical reasoning. | 16.4 | 20.7 |
FrontierMath-Tier-4-2025-07-01-Private FrontierMath Tier 4 (Jul 2025) · the most challenging tier of frontier mathematics, containing problems that push the absolute limits of AI mathematical reasoning. | 2.1 | 4.2 |
GPQA diamond Graduate-Level Google-Proof QA (Diamond set) · expert-crafted questions in physics, biology, and chemistry that are difficult even for domain PhDs. | 83.8 | 81.4 |
OTIS Mock AIME 2024-2025 OTIS Mock AIME 2024-2025 · simulated American Invitational Mathematics Examination problems testing advanced problem-solving skills. | 80.0 | 86.1 |
PostTrainBench | 13.9 | 17.3 |
SimpleBench SimpleBench · tests fundamental reasoning capabilities with straightforward problems designed to expose gaps in basic logical and spatial thinking. | 43.8 | 54.4 |
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. | 72.1 | 76.7 |
Terminal Bench Terminal-Bench 2.0 · evaluates AI agents on real terminal-based coding tasks · writing scripts, debugging, running tests, and managing projects entirely through command-line interaction. Tests both code quality and terminal fluency. Claude Opus 4.7 scores 69.4%, demonstrating significant agentic terminal competence. | 52.4 | 63.1 |
WeirdML WeirdML · tests models on unusual and adversarial machine learning tasks that require creative problem-solving beyond standard patterns. | 48.2 | 63.7 |
Pricing · per 1M tokens · projected $/mo at 10M tokens
| Model | Input | Output | Context | Projected $/mo |
|---|---|---|---|---|
| $0.60 | $1.92 | 203K tokens (~101 books) | $9.30 | |
| $5.00 | $25.00 | 200K tokens (~100 books) | $100.00 |
People also compared