Compare · ModelsLive · 2 picked · head to head
GLM 5.2 vs GLM 5.3 Flash
Side by side · benchmarks, pricing, and signals you can act on.
Winner summary
GLM 5.2 wins on 8/16 benchmarks
GLM 5.2 wins 8 of 16 shared benchmarks. Leads in speed · knowledge · math.
Category leads
speed·GLM 5.2agentic·GLM 5.3 Flashreasoning·GLM 5.3 Flasharena·GLM 5.3 Flashknowledge·GLM 5.2coding·GLM 5.3 Flashmath·GLM 5.2general·GLM 5.2
Hype vs Reality
Attention vs performance
GLM 5.2
#75 by perf·#3 by attention
GLM 5.3 Flash
#151 by perf·#3 by attention
Best value
GLM 5.3 Flash
9.2x better value than GLM 5.2
GLM 5.2
16.0 pts/$
$3.51/M
GLM 5.3 Flash
146.5 pts/$
$0.33/M
Vendor risk
Who is behind the model
z-ai
private · undisclosed
z-ai
private · undisclosed
Head to head
16 benchmarks · 2 models
GLM 5.2GLM 5.3 Flash
Artificial Analysis · Quality Index
GLM 5.2 leads by +9.3
GLM 5.2
51.1
GLM 5.3 Flash
41.8
APEX-Agents
GLM 5.3 Flash leads by +7.6
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
GLM 5.3 Flash
52.8
ARC-AGI
GLM 5.3 Flash leads by +14.0
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
GLM 5.3 Flash
91.0
ARC-AGI-2
GLM 5.3 Flash leads by +43.0
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
GLM 5.3 Flash
65.8
Chatbot Arena Elo · Coding
GLM 5.3 Flash leads by +22.3
GLM 5.2
1593.3
GLM 5.3 Flash
1615.6
Chatbot Arena Elo · Overall
GLM 5.3 Flash leads by +2.0
GLM 5.2
1471.0
GLM 5.3 Flash
1473.0
Chess Puzzles
GLM 5.2 leads by +1.1
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
GLM 5.3 Flash
16.9
Deepswe
GLM 5.3 Flash leads by +19.6
GLM 5.2
43.8
GLM 5.3 Flash
63.4
Frontiercode
GLM 5.3 Flash leads by +7.3
GLM 5.2
24.5
GLM 5.3 Flash
31.8
FrontierMath-Tier-4-v2-Private
GLM 5.2 leads by +12.2
GLM 5.2
29.3
GLM 5.3 Flash
17.1
FrontierMath-Tiers-1-3-v2-Private
GLM 5.2 leads by +3.4
GLM 5.2
59.2
GLM 5.3 Flash
55.8
GPQA diamond
GLM 5.2 leads by +2.3
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
GLM 5.3 Flash
86.9
Mystery Game Puzzles
GLM 5.2 leads by +10.8
GLM 5.2
10.8
GLM 5.3 Flash
0.0
OTIS Mock AIME 2024-2025
GLM 5.3 Flash leads by +7.5
OTIS Mock AIME 2024-2025 · simulated American Invitational Mathematics Examination problems testing advanced problem-solving skills.
GLM 5.2
86.4
GLM 5.3 Flash
93.9
Proofbench
GLM 5.2 leads by +14.0
GLM 5.2
35.0
GLM 5.3 Flash
21.0
Surface Evolver Bench
GLM 5.2 leads by +3.1
GLM 5.2
55.6
GLM 5.3 Flash
52.5
Full benchmark table
| Benchmark | GLM 5.2 | GLM 5.3 Flash |
|---|---|---|
Artificial Analysis · Quality Index | 51.1 | 41.8 |
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 | 52.8 |
ARC-AGI ARC-AGI · the original Abstraction and Reasoning Corpus, testing whether AI can solve novel visual pattern recognition tasks without memorization. | 77.0 | 91.0 |
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. | 22.8 | 65.8 |
Chatbot Arena Elo · Coding | 1593.3 | 1615.6 |
Chatbot Arena Elo · Overall | 1471.0 | 1473.0 |
Chess Puzzles Chess Puzzles · tests strategic and tactical reasoning by having models solve chess puzzle positions, evaluating lookahead and pattern recognition abilities. | 17.9 | 16.9 |
Deepswe | 43.8 | 63.4 |
Frontiercode | 24.5 | 31.8 |
FrontierMath-Tier-4-v2-Private | 29.3 | 17.1 |
FrontierMath-Tiers-1-3-v2-Private | 59.2 | 55.8 |
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 | 86.9 |
Mystery Game Puzzles | 10.8 | 0.0 |
OTIS Mock AIME 2024-2025 OTIS Mock AIME 2024-2025 · simulated American Invitational Mathematics Examination problems testing advanced problem-solving skills. | 86.4 | 93.9 |
Proofbench | 35.0 | 21.0 |
Surface Evolver Bench | 55.6 | 52.5 |
Pricing · per 1M tokens · projected $/mo at 10M tokens
| Model | Input | Output | Context | Projected $/mo |
|---|---|---|---|---|
| $0.02 | $7.00 | 1.0M tokens (~524 books) | $17.65 | |
| $0.15 | $0.50 | 1.0M tokens (~524 books) | $2.38 |