Compare / GLM-5.2 vs GLM-5.3

GLM-5.2 vs GLM-5.3

Which is cheaper, which is more capable, and which does better at each kind of work, from 22 results on 4 sources that measured both models.

Z.aiGLM-5.2 Released 16 Jun 2026 · $1.40 in · $4.40 out
Z.aiGLM-5.3 Released 14 Aug 2026 · $1.40 in · $4.40 out
GLM-5.2 wins 111 too close to callGLM-5.3 wins 10

Of the 22 results both models have, the better value on each. Results whose reported ranges overlap are too close to call.

Results as of 16 November 2026 (catalogue release 2026-11-16-49bdfa045dee).

GLM-5.2 or GLM-5.3? The short answer

GLM-5.3 is the newer version at about the same price, and it scores higher (Intelligence Index 44.8 against 33.7): upgrade.

  • They cost about the same: $2.15 and $2.15 per million tokens (three input to one output).
  • GLM-5.3 scores higher on the Artificial Analysis Intelligence Index: 44.8 against 33.7.
  • GLM-5.3 is newer: released 14 Aug 2026, against 16 Jun 2026 for GLM-5.2.
  • On third-party benchmarks, on the results both models have: GLM-5.3 does better for agents and tool use, reasoning and knowledge and sticking to the facts.

Head to head

33.7
Intelligence Index
44.8
$2.15
Price per million tokens, blendedlower is better
$2.15
1,048,576
Context window, tokens
1,048,576

Every fact side by side

GLM-5.2GLM-5.3
Price per million tokens$1.40 in · $4.40 out$1.40 in · $4.40 out
Blended (3 in : 1 out)$2.15$2.15
Cheapest host, blended$0.87 (DeepInfra, FP4)$0.40 (Wafer, FP4)
Cached input, per million$0.26 cached input$0.26 cached input
Intelligence Index33.744.8
Spring Prompt overall–35
Output speed, tokens a second–87
Context window1,048,576 tokens1,048,576 tokens
Released16 Jun 202614 Aug 2026 (newer)
WeightsOpenOpen
Developer based inChinaChina

Prices are the developer's list price where we have read it, otherwise the typical host on OpenRouter; see every model's price. Intelligence and speed from Artificial Analysis.

Which is better for what

Each model's rank among the models on sale today, for each kind of work, from our own benchmarks and licensed sources. Further right is better; the better rank is in green.

Marketing planningPlanning a year of ad spend without overspending not measured 29 of 32
Agents and tool useMulti-step tasks with tools: support desks, coding agents, function calls 47 of 159 15 of 159
Professional workReal tasks from banking, consulting and law, business documents and freelance projects not measured 15 of 51
Reasoning and knowledgeHard questions across science, maths and general knowledge 50 of 187 31 of 187
WritingWhat people prefer in blind comparisons, and judged writing quality 27 of 169 24 of 169
Sticking to the factsSummarising without inventing things, and factual answers 64 of 139 51 of 139
Speed under pressureGood decisions against a real clock (fast chess) not measured 33 of 35

Every shared result

Each model's best result on every measure both have, grouped by benchmark, with the setting that produced it. Under each measure, every model that benchmark has measured is a grey tick, best to the right, so you can see where the two sit in the field. The headline measures of each benchmark come first; the rest are under "Show all". Green marks the better value; where a source reports ranges that overlap, neither is marked.

MeasureGLM-5.2GLM-5.3
Arena (formerly LMArena) reported by Arena (formerly LMArena) · 3 results
Overall · TextArena rating, higher is better 27 · 24 of 387 measured 1,476max reasoning 1,478max reasoning
Overall · Text factualityArena rating, higher is better 36 · 29 of 163 measured 1,463max reasoning 1,467max reasoning
Business, management and finance · TextArena rating, higher is better 41 · 26 of 380 measured 1,465max reasoning 1,475max reasoning
Artificial Analysis reported by Artificial Analysis · 3 results
Artificial Analysis Intelligence Indexindex score, higher is better 46 · 17 of 249 measured 33.7max reasoning 44.8max reasoning
GPQA Diamond% of questions, higher is better 49 · 26 of 232 measured 89.5%max reasoning 91.7%max reasoning
Humanity's Last Exam% of questions, higher is better 41 · 38 of 250 measured 41.1%max reasoning 42.3%max reasoning
SimpleQA Verified (Epoch AI) reported by SimpleQA Verified (Epoch AI) · 1 result
Correct answers% of questions, higher is better 51 · 42 of 76 measured 34.2%max reasoning 41.0%max reasoning
Vending-Bench 2 reported by Vending-Bench 2 · 1 result
Money after a yearUS dollars, higher is better 7 · 8 of 58 measured $8,314 $8,164
Show all 22 results (14 more)
MeasureGLM-5.2GLM-5.3
Arena (formerly LMArena) reported by Arena (formerly LMArena) · 8 results
Confirmed task success · AgentIPS effect estimate, higher is better 19 · 16 of 48 measured 0.05max reasoning, arena agent 0.05max reasoning, arena agent
Creative writing · TextArena rating, higher is better 26 · 28 of 385 measured 1,454max reasoning 1,454max reasoning
Expert prompts · TextArena rating, higher is better 35 · 16 of 338 measured 1,502max reasoning 1,518max reasoning
Instruction following · TextArena rating, higher is better 31 · 21 of 387 measured 1,470max reasoning 1,476max reasoning
Praise over complaint · AgentIPS effect estimate, higher is better 12 · 18 of 48 measured 0.10max reasoning, arena agent 0.05max reasoning, arena agent
Steerability · AgentIPS effect estimate, higher is better 22 · 26 of 48 measured 0.01max reasoning, arena agent 0.01max reasoning, arena agent
Tool grounding · AgentIPS effect estimate, higher is better 9 · 12 of 48 measured 0.00max reasoning, arena agent 0.00max reasoning, arena agent
Writing, literature and language · TextArena rating, higher is better 31 · 23 of 386 measured 1,459max reasoning 1,466max reasoning
Artificial Analysis reported by Artificial Analysis · 6 results
Terminal-Bench 2.1% of tasks, higher is better 30 · 17 of 131 measured 77.9%max reasoning 83.9%max reasoning
Terminal-Bench 4.0% of tasks, higher is better 59 · 13 of 120 measured 1.0%max reasoning 41.9%max reasoning
τ-bench banking% of tasks, higher is better 27 · 4 of 121 measured 34.6%max reasoning 50.3%max reasoning
Artificial Analysis Coding Indexindex score, higher is better 30 · 14 of 130 measured 68.8max reasoning 74.8max reasoning
Long-context reasoning (AA-LCR)% of questions, higher is better 59 · 47 of 241 measured 78.3%max reasoning 79.7%max reasoning
SciCode% of problems, higher is better 48 · 12 of 121 measured 51.2%max reasoning 59.0%max reasoning

Quick answers

Is GLM-5.2 better than GLM-5.3?

GLM-5.3 is the newer version at about the same price, and it scores higher (Intelligence Index 44.8 against 33.7): upgrade.

Which is cheaper, GLM-5.2 or GLM-5.3?

They cost about the same: $2.15 and $2.15 per million tokens (three input to one output).

Which is better for agents and tool use, GLM-5.2 or GLM-5.3?

GLM-5.3 ranks 15 of 159 models on sale for multi-step tasks with tools: support desks, coding agents, function calls, against 47 for GLM-5.2.

Which is better for reasoning and knowledge, GLM-5.2 or GLM-5.3?

GLM-5.3 ranks 31 of 187 models on sale for hard questions across science, maths and general knowledge, against 50 for GLM-5.2.

Which is better for writing, GLM-5.2 or GLM-5.3?

GLM-5.3 ranks 24 of 169 models on sale for what people prefer in blind comparisons, and judged writing quality, against 27 for GLM-5.2.

Which is better for sticking to the facts, GLM-5.2 or GLM-5.3?

GLM-5.3 ranks 51 of 139 models on sale for summarising without inventing things, and factual answers, against 64 for GLM-5.2.

Run GLM-5.2 and GLM-5.3 on your own prompt.

Benchmarks aren't your task. Test both side by side on your prompt, with every answer checked against your rules and the cost of each one.