GLM 5.2 vs GLM 5.3

GLM 5.3 leads llmrun Score 130.1 to 122.9; neither fits 24 GB.

Model A · Z.ai

GLM 5.2

122.9llmrun Score

753.3B · 41.8B active · 1M context · MIT · Jun 2026

ChatFunctions

Model B · Z.ai

GLM 5.3

130.1llmrun Score

753.3B · 41.8B active · 1M context · Other · Aug 2026

ChatFunctions

Which should you choose?

Choose GLM 5.2 if you want…

More permissive license
MIT vs Other

Choose GLM 5.3 if you want…

Higher llmrun Score
130.1 vs 122.9
Stronger at reasoning
130.1 vs 119.8
Stronger at coding
130.1 vs 120.2
Ahead on the benchmarks
14 of 21 shared boards

On your hardware

Each model at its default quant (Q4_K_M), a smaller one when it must. Estimates, ±20%.

MemoryGLM 5.2GLM 5.3

8 GB

RTX 4060 · 8 GB tier

GLM 5.2

Doesn't fit

smallest quant needs 208.9 GB

GLM 5.3

Doesn't fit

smallest quant needs 256 GB

12 GB

RTX 3060 12GB · 12 GB tier

GLM 5.2

Doesn't fit

smallest quant needs 208.9 GB

GLM 5.3

Doesn't fit

smallest quant needs 256 GB

16 GB

RTX 5060 Ti 16GB · 16 GB tier

GLM 5.2

Doesn't fit

smallest quant needs 208.9 GB

GLM 5.3

Doesn't fit

smallest quant needs 256 GB

24 GBreference

RTX 4090 · 24 GB tier

GLM 5.2

Doesn't fit

smallest quant needs 208.9 GB

GLM 5.3

Doesn't fit

smallest quant needs 256 GB

32 GB

RTX 5090 · 32 GB tier

GLM 5.2

Doesn't fit

smallest quant needs 208.9 GB

GLM 5.3

Doesn't fit

smallest quant needs 256 GB

48 GB

RTX A6000 · 48 GB tier

GLM 5.2

Doesn't fit

smallest quant needs 208.9 GB

GLM 5.3

Doesn't fit

smallest quant needs 256 GB

Mac 32 GB

Mac Mini M4 · Macs

GLM 5.2

Doesn't fit

smallest quant needs 208.9 GB

GLM 5.3

Doesn't fit

smallest quant needs 256 GB

Mac 64 GB

Mac Studio M4 Max · Macs

GLM 5.2

Doesn't fit

smallest quant needs 208.9 GB

GLM 5.3

Doesn't fit

smallest quant needs 256 GB

Mac 128 GB

Mac Studio M4 Max · Macs

GLM 5.2

Doesn't fit

smallest quant needs 208.9 GB

GLM 5.3

Doesn't fit

smallest quant needs 256 GB

Reference GPU: RTX 4090. Tinted cell = the better pick on that machine (fits, more headroom, or at least 10% faster).

llmrun Score

with its 90% range · bars share one scale

GLM 5.2 GLM 5.3

Overall

GLM 5.2

122.9121–125

GLM 5.3

130.1128–132

Coding

GLM 5.2

120.2115–125

GLM 5.3

130.1124–136

Agents

GLM 5.2

128.2123–133

GLM 5.3

133.4130–137

Math

GLM 5.2

125.7123–128

GLM 5.3

128.6126–131

Science

GLM 5.2

128.5117–141

GLM 5.3

130.8119–143

Reasoning

GLM 5.2

119.8117–123

GLM 5.3

130.1124–136

GLM 5.2 · #10 of 100 open models · #8 of 96 that fit ≤512 GB

GLM 5.3 · #4 of 100 open models · #3 of 96 that fit ≤512 GB

Benchmarks

21 shared boards: GLM 5.2 leads 6, GLM 5.3 leads 14 · Benchmarks imported Oct 2026

BoardGLM 5.2GLM 5.3

GPQA Diamond

GLM 5.2

91.9%#2/98 open

IndependentEpoch AI · Jun 2026

GLM 5.3

90.9%#7/98 open

IndependentEpoch AI · Aug 2026

LiveBench Coding

GLM 5.2

79.7#3/19 open

IndependentLiveBench · Jun 2026

GLM 5.3

79.0#4/19 open

IndependentLiveBench · Jun 2026

LiveBench Reasoning

GLM 5.2

78.6#13/19 open

IndependentLiveBench · Jun 2026

GLM 5.3

85.8#7/19 open

IndependentLiveBench · Jun 2026

AIME 2024/2025

GLM 5.2

86.4%#21/86 open

IndependentEpoch AI · Jun 2026

GLM 5.3

91.1%#13/86 open

IndependentEpoch AI · Aug 2026

FrontierMath

GLM 5.2

59.2%#6/17 open

IndependentEpoch AI · Jun 2026

GLM 5.3

68.8%#3/17 open

IndependentEpoch AI · Aug 2026

LiveBench Math

GLM 5.2

89.8#5/19 open

IndependentLiveBench · Jun 2026

GLM 5.3

87.9#8/19 open

IndependentLiveBench · Jun 2026

SciCode

GLM 5.2

51.2%#8/66 open

Lab-reportedEpoch AI · Jun 2026

GLM 5.3

59.0%#3/66 open

Lab-reportedEpoch AI · Aug 2026

SimpleQA

GLM 5.2

34.2%#12/17 open

IndependentEpoch AI · Jun 2026

GLM 5.3

41.0%#6/17 open

IndependentEpoch AI · Aug 2026

FrontierMath Tier 4

GLM 5.2

29.3%#3/13 open

IndependentEpoch AI · Jun 2026

GLM 5.3

29.3%#4/13 open

IndependentEpoch AI · Aug 2026

WeirdML

GLM 5.2

70.1%#3/46 open

IndependentEpoch AI · Jun 2026

GLM 5.3

75.4%#2/46 open

IndependentEpoch AI · Aug 2026

27 more boards (other shared and single-model boards)

DeepSWE

GLM 5.2

43.8%#4/5 open

Lab-reportedEpoch AI · Jun 2026

GLM 5.3

69.0%#1/5 open

Lab-reportedEpoch AI · Aug 2026

FrontierCode

GLM 5.2

24.5%#7/12 open

Lab-reportedEpoch AI · Jun 2026

GLM 5.3

40.1%#3/12 open

Lab-reportedEpoch AI · Aug 2026

ALE-Bench

GLM 5.2

1047.0#9/43 open

Lab-reportedEpoch AI · Jun 2026

GLM 5.3

1317.4#3/43 open

Lab-reportedEpoch AI · Aug 2026

APEX-Agents

GLM 5.2

45.2%#7/16 open

Lab-reportedEpoch AI · Jun 2026

GLM 5.3

56.6%#2/16 open

Lab-reportedEpoch AI · Aug 2026

Chess Puzzles

GLM 5.2

22.0%#6/14 open

IndependentEpoch AI · Jun 2026

GLM 5.3

37.0%#4/14 open

IndependentEpoch AI · Aug 2026

CritPt

GLM 5.2

20.9%#3/66 open

Lab-reportedEpoch AI · Jun 2026

GLM 5.3

19.1%#4/66 open

Lab-reportedEpoch AI · Aug 2026

DTBench

GLM 5.2

93.6%#2/68 open

Lab-reportedEpoch AI · Jun 2026

GLM 5.3

87.7%#11/68 open

Lab-reportedEpoch AI · Aug 2026

LMCA

GLM 5.2

45.8%#4/54 open

Lab-reportedEpoch AI · Jun 2026

GLM 5.3

55.5%#1/54 open

Lab-reportedEpoch AI · Aug 2026

Mystery Game Puzzles

GLM 5.2

19.0%#8/18 open

IndependentEpoch AI · Jun 2026

GLM 5.3

33.0%#4/18 open

IndependentEpoch AI · Aug 2026

ProofBench

GLM 5.2

35.0%#9/26 open

Lab-reportedEpoch AI · Jun 2026

GLM 5.3

49.0%#8/26 open

Lab-reportedEpoch AI · Aug 2026

Vending-Bench 2

GLM 5.2

8313.8#1/17 open

Lab-reportedEpoch AI · Jun 2026

GLM 5.3

8163.6#2/17 open

Lab-reportedEpoch AI · Aug 2026

LMArena Text

GLM 5.2

1475.9#4/187 open

ArenaLMArena · Oct 2026

GLM 5.3

1478.5#3/187 open

ArenaLMArena · Oct 2026

LMArena WebDev

GLM 5.2

1604.6#9/52 open

ArenaLMArena · Oct 2026

GLM 5.3

1621.1#6/52 open

ArenaLMArena · Oct 2026

Only GLM 5.2 (5)

Only GLM 5.3 (9)

VRAM by quantization

Memory to load the model and 2K context; below it, the total at 32K and 128K context. Lighter side in bold.

QuantGLM 5.2GLM 5.3

Q2_K

GLM 5.2

324 GB~320.2 GB file

32K 323 GB · 128K 332 GB

GLM 5.3

324 GB~320.2 GB file

32K 323 GB · 128K 332 GB

Q3_K_M

GLM 5.2

371 GB342.7 GB file

32K 370 GB · 128K 379 GB

GLM 5.3

371 GB~367.3 GB file

32K 370 GB · 128K 379 GB

Q4_K_M

GLM 5.2

456 GB465.8 GB file

32K 455 GB · 128K 464 GB

GLM 5.3

456 GB~452.0 GB file

32K 455 GB · 128K 464 GB

Q5_K_M

GLM 5.2

541 GB560.8 GB file

32K 540 GB · 128K 549 GB

GLM 5.3

541 GB~536.8 GB file

32K 540 GB · 128K 549 GB

Q6_K

GLM 5.2

626 GB625.9 GB file

32K 625 GB · 128K 634 GB

GLM 5.3

626 GB~621.5 GB file

32K 625 GB · 128K 634 GB

Q8_0

GLM 5.2

758 GB801.4 GB file

32K 757 GB · 128K 765 GB

GLM 5.3

758 GB801.4 GB file

32K 757 GB · 128K 765 GB

10 more quants

IQ2_XXS

GLM 5.2

211 GB238.5 GB file

32K 210 GB · 128K 219 GB

GLM 5.3

not listed

IQ2_XS

GLM 5.2

230 GB230.0 GB file

32K 229 GB · 128K 238 GB

GLM 5.3

not listed

IQ2_M

GLM 5.2

258 GB238.6 GB file

32K 257 GB · 128K 266 GB

GLM 5.3

258 GB238.6 GB file

32K 257 GB · 128K 266 GB

IQ3_XXS

GLM 5.2

296 GB281.7 GB file

32K 295 GB · 128K 304 GB

GLM 5.3

296 GB281.7 GB file

32K 295 GB · 128K 304 GB

IQ3_S

GLM 5.2

324 GB308.6 GB file

32K 323 GB · 128K 332 GB

GLM 5.3

not listed

IQ4_XS

GLM 5.2

409 GB365.3 GB file

32K 408 GB · 128K 417 GB

GLM 5.3

409 GB365.3 GB file

32K 408 GB · 128K 417 GB

IQ4_NL

GLM 5.2

428 GB372.7 GB file

32K 427 GB · 128K 436 GB

GLM 5.3

not listed

Q4_K_S

GLM 5.2

428 GB436.4 GB file

32K 427 GB · 128K 436 GB

GLM 5.3

not listed

Q5_K_S

GLM 5.2

522 GB527.3 GB file

32K 521 GB · 128K 530 GB

GLM 5.3

not listed

BF16

GLM 5.2

1511 GB1508.0 GB file

32K 1510 GB · 128K 1519 GB

GLM 5.3

1511 GB1508.0 GB file

32K 1510 GB · 128K 1519 GB

How to run each

GLM 5.2

Ollama · Q4_K_M

$ollama run hf.co/unsloth/GLM-5.2-GGUF:Q4_K_M

All ways to run GLM 5.2 →

GLM 5.3

Ollama · Q4_K_M

$ollama run hf.co/unsloth/GLM-5.3-GGUF

All ways to run GLM 5.3 →

Specifications

Specifications of GLM 5.2 and GLM 5.3
Spec	GLM 5.2	GLM 5.3
Publisher	Z.ai	Z.ai
Parameters	753.3B	753.3B
Active parameters	42B per token	42B per token
Architecture	GlmMoeDsaForCausalLM	GlmMoeDsaForCausalLM
Context	1M tokens	1M tokens
Modalities	Chat, Function-calling	Chat, Function-calling
License	MIT	Other
Released	Jun 2026	Aug 2026

Frequently Asked Questions

Which is better, GLM 5.2 or GLM 5.3?

GLM 5.3 leads llmrun Score 130.1 to 122.9; neither fits 24 GB. Choose GLM 5.2 for more permissive license; choose GLM 5.3 for higher llmrun Score and stronger at reasoning.

Which needs less VRAM, GLM 5.2 or GLM 5.3?

At Q4_K_M, GLM 5.2 requires 456.2 GB while GLM 5.3 needs 456.2 GB, making GLM 5.2 the easier model to run on your local setup.

Can I run GLM 5.2 or GLM 5.3 on a 16 GB GPU?

GLM 5.2 does not fit (smallest quant needs 208.9 GB); GLM 5.3 does not fit (smallest quant needs 256 GB).

Which has a longer context window, GLM 5.2 or GLM 5.3?

GLM 5.2 supports 1,048,576 tokens and GLM 5.3 supports 1,048,576 tokens.