Compare · ModelsLive · 3 picked · head to head
GLM 5.2 vs Kimi K2.7 Code vs MiniMax M3
Side by side · benchmarks, pricing, and signals you can act on.
Winner summary
GLM 5.2 wins on 22/38 benchmarks
GLM 5.2 wins 22 of 38 shared benchmarks. Leads in knowledge · coding · language.
Category leads
speed·MiniMax M3knowledge·GLM 5.2coding·GLM 5.2reasoning·MiniMax M3language·GLM 5.2math·GLM 5.2general·GLM 5.2agentic·MiniMax M3arena·GLM 5.2
Hype vs Reality
Attention vs performance
GLM 5.2
#67 by perf·#3 by attention
Kimi K2.7 Code
#69 by perf·#17 by attention
MiniMax M3
#124 by perf·#11 by attention
Best value
MiniMax M3
2.2x better value than Kimi K2.7 Code
GLM 5.2
22.6 pts/$
$2.50/M
Kimi K2.7 Code
30.5 pts/$
$1.84/M
MiniMax M3
66.5 pts/$
$0.75/M
Vendor risk
Mixed exposure
One or more vendors flagged
z-ai
private · undisclosed
Moonshot AI
$18.0B·Tier 1
MiniMax
$4.0B·Tier 1
Head to head
38 benchmarks · 3 models
GLM 5.2Kimi K2.7 CodeMiniMax M3
Artificial Analysis · Agentic Index
GLM 5.2 leads by +7.7
Artificial Analysis Agentic Index · a composite score measuring how well a model performs in agentic workflows · multi-step tool use, planning, error recovery, and autonomous task completion. Aggregates results from multiple agentic benchmarks including SWE-bench, tool-use tests, and planning evaluations. The canonical single-number metric for "how good is this model as an agent?"
GLM 5.2
43.1
Kimi K2.7 Code
29.6
MiniMax M3
35.4
Artificial Analysis · Coding Index
GLM 5.2 leads by +8.0
Artificial Analysis Coding Index · a composite score that aggregates performance across multiple coding benchmarks into a single index. Tracks code generation quality, debugging ability, multi-language competence, and real-world software engineering tasks. Used by Artificial Analysis to rank model coding capability in a normalized, comparable format. Useful for developers choosing between models for coding-heavy workloads.
GLM 5.2
68.8
Kimi K2.7 Code
60.8
MiniMax M3
58.6
Artificial Analysis · Quality Index
GLM 5.2 leads by +21.9
GLM 5.2
51.1
Kimi K2.7 Code
25.8
MiniMax M3
29.2
Chess Puzzles
Chess Puzzles · tests strategic and tactical reasoning by having models solve chess puzzle positions, evaluating lookahead and pattern recognition abilities.
GLM 5.2
16.9
Kimi K2.7 Code
16.9
MiniMax M3
9.5
Frontiercode
Kimi K2.7 Code leads by +5.6
GLM 5.2
24.5
Kimi K2.7 Code
30.1
MiniMax M3
14.7
GPQA diamond
GLM 5.2 leads by +1.3
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
89.1
Kimi K2.7 Code
83.8
MiniMax M3
87.9
LiveBench · Agentic Coding
GLM 5.2 leads by +3.3
GLM 5.2
73.3
Kimi K2.7 Code
70.0
MiniMax M3
60.0
LiveBench · Coding
GLM 5.2 leads by +5.7
GLM 5.2
79.7
Kimi K2.7 Code
74.0
MiniMax M3
68.2
LiveBench · Data Analysis
MiniMax M3 leads by +2.4
GLM 5.2
73.7
Kimi K2.7 Code
62.7
MiniMax M3
76.2
LiveBench · If
GLM 5.2 leads by +4.8
GLM 5.2
62.3
Kimi K2.7 Code
56.3
MiniMax M3
57.5
LiveBench · Language
Kimi K2.7 Code leads by +1.1
GLM 5.2
76.2
Kimi K2.7 Code
77.9
MiniMax M3
76.8
LiveBench · Mathematics
GLM 5.2 leads by +10.2
GLM 5.2
89.8
Kimi K2.7 Code
79.6
MiniMax M3
77.0
LiveBench · Overall
GLM 5.2 leads by +4.3
GLM 5.2
76.2
Kimi K2.7 Code
71.9
MiniMax M3
70.0
LiveBench · Reasoning
Kimi K2.7 Code leads by +4.2
GLM 5.2
78.6
Kimi K2.7 Code
82.8
MiniMax M3
74.5
OTIS Mock AIME 2024-2025
Kimi K2.7 Code leads by +9.2
OTIS Mock AIME 2024-2025 · simulated American Invitational Mathematics Examination problems testing advanced problem-solving skills.
GLM 5.2
86.4
Kimi K2.7 Code
95.5
MiniMax M3
71.1
SimpleBench
GLM 5.2 leads by +1.1
SimpleBench · tests fundamental reasoning capabilities with straightforward problems designed to expose gaps in basic logical and spatial thinking.
GLM 5.2
50.6
Kimi K2.7 Code
49.5
MiniMax M3
35.0
Surface Evolver Bench
GLM 5.2 leads by +0.6
GLM 5.2
55.6
Kimi K2.7 Code
48.8
MiniMax M3
55.0
Artificial Analysis · CritPt
Kimi K2.7 Code leads by +6.3
Kimi K2.7 Code
10.0
MiniMax M3
3.7
Artificial Analysis · GDPval
MiniMax M3 leads by +11.0
Kimi K2.7 Code
26.3
MiniMax M3
37.3
Artificial Analysis · GPQA Diamond
MiniMax M3 leads by +3.3
Kimi K2.7 Code
89.6
MiniMax M3
92.9
Artificial Analysis · Humanity's Last Exam
MiniMax M3 leads by +4.0
Kimi K2.7 Code
35.0
MiniMax M3
39.0
Artificial Analysis · IFBench
MiniMax M3 leads by +19.8
Kimi K2.7 Code
63.1
MiniMax M3
82.9
Artificial Analysis · Long Context Reasoning
MiniMax M3 leads by +3.7
Kimi K2.7 Code
79.3
MiniMax M3
83.0
Artificial Analysis · SciCode
Kimi K2.7 Code leads by +0.7
Kimi K2.7 Code
47.8
MiniMax M3
47.1
Artificial Analysis · tau2-Bench Telecom
Kimi K2.7 Code leads by +1.2
Kimi K2.7 Code
90.1
MiniMax M3
88.9
Artificial Analysis · Terminal-Bench Hard
Kimi K2.7 Code leads by +2.3
Kimi K2.7 Code
44.7
MiniMax M3
42.4
APEX-Agents
MiniMax M3 leads by +0.1
APEX-Agents · evaluates AI agents on complex, multi-step tasks requiring planning, tool use, and autonomous decision-making in realistic environments.
Kimi K2.7 Code
37.6
MiniMax M3
37.7
Chatbot Arena Elo · Coding
GLM 5.2 leads by +111.2
GLM 5.2
1593.3
MiniMax M3
1482.0
Chatbot Arena Elo · Overall
GLM 5.2 leads by +30.9
GLM 5.2
1471.0
MiniMax M3
1440.1
Deepswe
GLM 5.2 leads by +13.3
GLM 5.2
43.8
Kimi K2.7 Code
30.5
Dtbench
GLM 5.2 leads by +24.5
GLM 5.2
89.3
MiniMax M3
64.9
FrontierMath-Tier-4-v2-Private
GLM 5.2 leads by +17.1
GLM 5.2
29.3
Kimi K2.7 Code
12.2
FrontierMath-Tiers-1-3-v2-Private
GLM 5.2 leads by +5.2
GLM 5.2
59.2
Kimi K2.7 Code
54.0
Lmca
GLM 5.2 leads by +14.3
GLM 5.2
53.9
MiniMax M3
39.6
Mystery Game Puzzles
GLM 5.2 leads by +10.8
GLM 5.2
10.8
MiniMax M3
0.0
Proofbench
GLM 5.2 leads by +17.0
GLM 5.2
35.0
MiniMax M3
18.0
SimpleQA Verified
Kimi K2.7 Code leads by +2.3
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
Kimi K2.7 Code
36.5
WeirdML
GLM 5.2 leads by +16.0
WeirdML · tests models on unusual and adversarial machine learning tasks that require creative problem-solving beyond standard patterns.
GLM 5.2
70.1
Kimi K2.7 Code
54.1
Full benchmark table
| Benchmark | GLM 5.2 | Kimi K2.7 Code | MiniMax M3 |
|---|---|---|---|
Artificial Analysis · Agentic Index Artificial Analysis Agentic Index · a composite score measuring how well a model performs in agentic workflows · multi-step tool use, planning, error recovery, and autonomous task completion. Aggregates results from multiple agentic benchmarks including SWE-bench, tool-use tests, and planning evaluations. The canonical single-number metric for "how good is this model as an agent?" | 43.1 | 29.6 | 35.4 |
Artificial Analysis · Coding Index Artificial Analysis Coding Index · a composite score that aggregates performance across multiple coding benchmarks into a single index. Tracks code generation quality, debugging ability, multi-language competence, and real-world software engineering tasks. Used by Artificial Analysis to rank model coding capability in a normalized, comparable format. Useful for developers choosing between models for coding-heavy workloads. | 68.8 | 60.8 | 58.6 |
Artificial Analysis · Quality Index | 51.1 | 25.8 | 29.2 |
Chess Puzzles Chess Puzzles · tests strategic and tactical reasoning by having models solve chess puzzle positions, evaluating lookahead and pattern recognition abilities. | 16.9 | 16.9 | 9.5 |
Frontiercode | 24.5 | 30.1 | 14.7 |
GPQA diamond Graduate-Level Google-Proof QA (Diamond set) · expert-crafted questions in physics, biology, and chemistry that are difficult even for domain PhDs. | 89.1 | 83.8 | 87.9 |
LiveBench · Agentic Coding | 73.3 | 70.0 | 60.0 |
LiveBench · Coding | 79.7 | 74.0 | 68.2 |
LiveBench · Data Analysis | 73.7 | 62.7 | 76.2 |
LiveBench · If | 62.3 | 56.3 | 57.5 |
LiveBench · Language | 76.2 | 77.9 | 76.8 |
LiveBench · Mathematics | 89.8 | 79.6 | 77.0 |
LiveBench · Overall | 76.2 | 71.9 | 70.0 |
LiveBench · Reasoning | 78.6 | 82.8 | 74.5 |
OTIS Mock AIME 2024-2025 OTIS Mock AIME 2024-2025 · simulated American Invitational Mathematics Examination problems testing advanced problem-solving skills. | 86.4 | 95.5 | 71.1 |
SimpleBench SimpleBench · tests fundamental reasoning capabilities with straightforward problems designed to expose gaps in basic logical and spatial thinking. | 50.6 | 49.5 | 35.0 |
Surface Evolver Bench | 55.6 | 48.8 | 55.0 |
Artificial Analysis · CritPt | — | 10.0 | 3.7 |
Artificial Analysis · GDPval | — | 26.3 | 37.3 |
Artificial Analysis · GPQA Diamond | — | 89.6 | 92.9 |
Artificial Analysis · Humanity's Last Exam | — | 35.0 | 39.0 |
Artificial Analysis · IFBench | — | 63.1 | 82.9 |
Artificial Analysis · Long Context Reasoning | — | 79.3 | 83.0 |
Artificial Analysis · SciCode | — | 47.8 | 47.1 |
Artificial Analysis · tau2-Bench Telecom | — | 90.1 | 88.9 |
Artificial Analysis · Terminal-Bench Hard | — | 44.7 | 42.4 |
APEX-Agents APEX-Agents · evaluates AI agents on complex, multi-step tasks requiring planning, tool use, and autonomous decision-making in realistic environments. | — | 37.6 | 37.7 |
Chatbot Arena Elo · Coding | 1593.3 | — | 1482.0 |
Chatbot Arena Elo · Overall | 1471.0 | — | 1440.1 |
Deepswe | 43.8 | 30.5 | — |
Dtbench | 89.3 | — | 64.9 |
FrontierMath-Tier-4-v2-Private | 29.3 | 12.2 | — |
FrontierMath-Tiers-1-3-v2-Private | 59.2 | 54.0 | — |
Lmca | 53.9 | — | 39.6 |
Mystery Game Puzzles | 10.8 | — | 0.0 |
Proofbench | 35.0 | — | 18.0 |
SimpleQA Verified SimpleQA Verified · short factual questions with verified answers, measuring factual accuracy and the tendency to hallucinate or provide incorrect information. | 34.2 | 36.5 | — |
WeirdML WeirdML · tests models on unusual and adversarial machine learning tasks that require creative problem-solving beyond standard patterns. | 70.1 | 54.1 | — |
Pricing · per 1M tokens · projected $/mo at 10M tokens
| Model | Input | Output | Context | Projected $/mo |
|---|---|---|---|---|
| $1.00 | $4.00 | 1.0M tokens (~524 books) | $17.50 | |
| $0.61 | $3.07 | 262K tokens (~131 books) | $12.26 | |
| $0.30 | $1.20 | 1.0M tokens (~524 books) | $5.25 |