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.
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
Every fact side by side
| GLM-5.2 | GLM-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 Index | 33.7 | 44.8 |
| Spring Prompt overall | – | 35 |
| Output speed, tokens a second | – | 87 |
| Context window | 1,048,576 tokens | 1,048,576 tokens |
| Released | 16 Jun 2026 | 14 Aug 2026 (newer) |
| Weights | Open | Open |
| Developer based in | China | China |
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.
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.
| Measure | GLM-5.2 | GLM-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)
| Measure | GLM-5.2 | GLM-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.