Compare · ModelsLive · 2 picked · head to head
Kimi K2.5 vs GLM 5
Side by side · benchmarks, pricing, and signals you can act on.
Winner summary
Kimi K2.5 wins on 11/19 benchmarks
Kimi K2.5 wins 11 of 19 shared benchmarks. Leads in reasoning · knowledge · math.
Category leads
agentic·GLM 5reasoning·Kimi K2.5knowledge·Kimi K2.5math·Kimi K2.5language·Kimi K2.5coding·GLM 5
Hype vs Reality
Attention vs performance
Kimi K2.5
#104 by perf·no signal
GLM 5
#79 by perf·#27 by attention
Best value
GLM 5
1.0x better value than Kimi K2.5
Kimi K2.5
43.3 pts/$
$1.20/M
GLM 5
44.5 pts/$
$1.26/M
Vendor risk
Who is behind the model
moonshotai
private · undisclosed
z-ai
private · undisclosed
Head to head
19 benchmarks · 2 models
Kimi K2.5GLM 5
APEX-Agents
GLM 5 leads by +2.8
APEX-Agents · evaluates AI agents on complex, multi-step tasks requiring planning, tool use, and autonomous decision-making in realistic environments.
Kimi K2.5
14.4
GLM 5
17.2
ARC-AGI
Kimi K2.5 leads by +20.7
ARC-AGI · the original Abstraction and Reasoning Corpus, testing whether AI can solve novel visual pattern recognition tasks without memorization.
Kimi K2.5
65.3
GLM 5
44.7
ARC-AGI-2
Kimi K2.5 leads by +7.0
ARC-AGI-2 · the second iteration of the Abstraction and Reasoning Corpus, testing novel pattern recognition and abstract reasoning without prior training data.
Kimi K2.5
11.8
GLM 5
4.9
Chess Puzzles
Kimi K2.5 leads by +2.0
Chess Puzzles · tests strategic and tactical reasoning by having models solve chess puzzle positions, evaluating lookahead and pattern recognition abilities.
Kimi K2.5
12.0
GLM 5
10.0
FrontierMath-2025-02-28-Private
Kimi K2.5 leads by +11.5
FrontierMath (Feb 2025) · original research-level math problems created by mathematicians, testing capabilities at the boundary of current AI mathematical reasoning.
Kimi K2.5
27.9
GLM 5
16.4
FrontierMath-Tier-4-2025-07-01-Private
Kimi K2.5 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.
Kimi K2.5
4.2
GLM 5
2.1
GPQA diamond
GLM 5 leads by +0.3
Graduate-Level Google-Proof QA (Diamond set) · expert-crafted questions in physics, biology, and chemistry that are difficult even for domain PhDs.
Kimi K2.5
83.5
GLM 5
83.8
OpenCompass · AIME2025
GLM 5 leads by +3.9
Kimi K2.5
91.9
GLM 5
95.8
OpenCompass · GPQA-Diamond
Kimi K2.5 leads by +2.8
Kimi K2.5
88.1
GLM 5
85.3
OpenCompass · HLE
Kimi K2.5 leads by +0.5
Kimi K2.5
28.6
GLM 5
28.1
OpenCompass · IFEval
Kimi K2.5 leads by +0.7
Kimi K2.5
93.9
GLM 5
93.2
OpenCompass · LiveCodeBenchV6
GLM 5 leads by +5.6
Kimi K2.5
80.6
GLM 5
86.2
OpenCompass · MMLU-Pro
Kimi K2.5 leads by +1.0
Kimi K2.5
86.2
GLM 5
85.2
OTIS Mock AIME 2024-2025
Kimi K2.5 leads by +12.2
OTIS Mock AIME 2024-2025 · simulated American Invitational Mathematics Examination problems testing advanced problem-solving skills.
Kimi K2.5
92.2
GLM 5
80.0
PostTrainBench
GLM 5 leads by +3.6
Kimi K2.5
10.3
GLM 5
13.9
SimpleBench
GLM 5 leads by +7.7
SimpleBench · tests fundamental reasoning capabilities with straightforward problems designed to expose gaps in basic logical and spatial thinking.
Kimi K2.5
36.2
GLM 5
43.8
SWE-Bench verified
Kimi K2.5 leads by +1.7
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.
Kimi K2.5
73.8
GLM 5
72.1
Terminal Bench
GLM 5 leads by +9.2
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.
Kimi K2.5
43.2
GLM 5
52.4
WeirdML
GLM 5 leads by +2.6
WeirdML · tests models on unusual and adversarial machine learning tasks that require creative problem-solving beyond standard patterns.
Kimi K2.5
45.6
GLM 5
48.2
Full benchmark table
| Benchmark | Kimi K2.5 | GLM 5 |
|---|---|---|
APEX-Agents APEX-Agents · evaluates AI agents on complex, multi-step tasks requiring planning, tool use, and autonomous decision-making in realistic environments. | 14.4 | 17.2 |
ARC-AGI ARC-AGI · the original Abstraction and Reasoning Corpus, testing whether AI can solve novel visual pattern recognition tasks without memorization. | 65.3 | 44.7 |
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. | 11.8 | 4.9 |
Chess Puzzles Chess Puzzles · tests strategic and tactical reasoning by having models solve chess puzzle positions, evaluating lookahead and pattern recognition abilities. | 12.0 | 10.0 |
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. | 27.9 | 16.4 |
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. | 4.2 | 2.1 |
GPQA diamond Graduate-Level Google-Proof QA (Diamond set) · expert-crafted questions in physics, biology, and chemistry that are difficult even for domain PhDs. | 83.5 | 83.8 |
OpenCompass · AIME2025 | 91.9 | 95.8 |
OpenCompass · GPQA-Diamond | 88.1 | 85.3 |
OpenCompass · HLE | 28.6 | 28.1 |
OpenCompass · IFEval | 93.9 | 93.2 |
OpenCompass · LiveCodeBenchV6 | 80.6 | 86.2 |
OpenCompass · MMLU-Pro | 86.2 | 85.2 |
OTIS Mock AIME 2024-2025 OTIS Mock AIME 2024-2025 · simulated American Invitational Mathematics Examination problems testing advanced problem-solving skills. | 92.2 | 80.0 |
PostTrainBench | 10.3 | 13.9 |
SimpleBench SimpleBench · tests fundamental reasoning capabilities with straightforward problems designed to expose gaps in basic logical and spatial thinking. | 36.2 | 43.8 |
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. | 73.8 | 72.1 |
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. | 43.2 | 52.4 |
WeirdML WeirdML · tests models on unusual and adversarial machine learning tasks that require creative problem-solving beyond standard patterns. | 45.6 | 48.2 |
Pricing · per 1M tokens · projected $/mo at 10M tokens