GLM 5.3 vs GPT-5.2
GLM 5.3 leads GPT-5.2 on coding, 70.6 to 55.9 on the Vibe Coding Index, and GLM 5.3 costs 3.2x less per output token ($4.40 vs $14.00 per million). Updated October 22, 2026; 66 benchmarks have a result for both.
GLM 5.3 or GPT-5.2: which to pick
Pick GLM 5.3 for coding: it scores 70.6 on the Vibe Coding Index against 55.9 for GPT-5.2.
Pick GPT-5.2 for work measured by LiveCodeBench (Vals AI) or GPQA Diamond, where it leads.
- Shared results
- 66
- GLM 5.3 only
- 34
- GPT-5.2 only
- 48
- Like-for-like categories
- 8 of 9
Only benchmarks that tested both models count toward a pick; a result only one model has is shown but not compared.
Which one for your work
Each card is decided by one number both models carry. Where only one has it, the card says so and makes no call.
Coding
Vibe Coding Index, six public coding benchmarks
GLM 5.3
GLM 5.3 scores 70.6 against 55.9.
6 and 3 benchmarks behind the scores
Intelligence
AA Intelligence Index
GLM 5.3
GLM 5.3 leads, 44.8 against 30.4.
Same board, both measured
Agentic work
AA Agentic Index
No call
Only GLM 5.3 has a AA Agentic Index result, 53.1.
Directional only
Reasoning
GPQA Diamond
GPT-5.2
GPT-5.2 leads, 91.7% against 88.1%.
Same board, both measured
Math
AIME
No call
Only GPT-5.2 has a AIME result, 96.9%.
Directional only
Knowledge
MMLU Pro
GLM 5.3
GLM 5.3 leads, 86.8% against 86.2%.
Same board, both measured
Web design
Design Arena: websites
GLM 5.3
GLM 5.3 leads, 1295 against 1195.
Same board, both measured
General
Vals Index
No call
Only GLM 5.3 has a Vals Index result, 59.7%.
Directional only
Long documents
Listed context window
GLM 5.3
GLM 5.3 takes 1.05M tokens against 400K.
Listed by OpenRouter
Everyday cost
1K input + 500 output tokens
GLM 5.3
GLM 5.3 $0.0036, GPT-5.2 $0.0088 per turn.
List prices
Cost per solved task
One SWE-bench Verified task, solved
GLM 5.3
GLM 5.3 $0.36, GPT-5.2 $1.91.
Published attempt cost and pass rate
At a glance
- Coding index
- GLM 5.3 70.6 from 6 benchmarks, GPT-5.2 55.9 from 3 benchmarks.
- Widest category
- Agentic work, 20.6 points apart on average over 2 shared benchmarks.
- Widest single gap
- ProofBench: GLM 5.3 49.0% against GPT-5.2 15.0%, 34.0 points apart on a 0 to 100 scale.
- Output price
- GLM 5.3 $4.40 against GPT-5.2 $14.00 per million output tokens, 3.2x apart.
- Context window
- GLM 5.3 takes 1.05M tokens, GPT-5.2 400K.
Benchmark by benchmark
First the score: boards each model leads, per category, where both have a result. Then every board, a category to a fold; bars are each result on a 0 to 100 scale, so compare the two in a row, not one row with another.
| Category | GLM 5.3 | Bars | GPT-5.2 | Ties |
|---|---|---|---|---|
| Coding | 8 | 1 | ||
| Intelligence | 2 | 0 | ||
| Agentic work | 2 | 0 | ||
| Reasoning | 6 | 3 | ||
| Math | 4 | 2 | ||
| Knowledge | 7 | 2 | ||
| Web design | 8 | 0 | ||
| General | 19 | 2 | ||
| All shared boards | 56 | 10 |
CodingLike-for-likeSix of these boards feed the Vibe Coding Index81
| Benchmark | GLM 5.3 | GPT-5.2 | Leader |
|---|---|---|---|
| SWE-bench Verified | Not measured | 72.8% | Only GPT-5.2 measured |
| SWE-bench Multilingual | Not measured | 66.7% | Only GPT-5.2 measured |
| LiveBench coding | 82.8% | Not measured | Only GLM 5.3 measured |
| LMArena WebDev | 1621 | 1417 | GLM 5.3 |
| AA Coding Index | 74.8 | Not measured | Only GLM 5.3 measured |
| LiveCodeBench (Vals AI) | 80.5% | 85.4% | GPT-5.2 |
| SWE-bench Verified (Vals AI) | 95.4% | 75.8% | GLM 5.3 |
| Terminal-Bench 4 (Vals AI) | 38.9% | Not measured | Only GLM 5.3 measured |
| Vibe Code Bench (Vals AI) | 78.1% | 53.5% | GLM 5.3 |
| Code migration (Vals AI) | 44.2% | Not measured | Only GLM 5.3 measured |
| ProgramBench (Vals AI) | 1.5% | Not measured | Only GLM 5.3 measured |
| IOI (Vals AI) | 68.4% | Not measured | Only GLM 5.3 measured |
| Terminal-Bench 2.1 (Vals AI) | 71.5% | Not measured | Only GLM 5.3 measured |
| Vibe Code Bench 1-100 (Vals AI) | 20.0% | Not measured | Only GLM 5.3 measured |
| SkillsBench (Vals AI) | 47.5% | Not measured | Only GLM 5.3 measured |
| CyberBench (Vals AI) | 72.1% | Not measured | Only GLM 5.3 measured |
| LiveCodeBench (Artificial Analysis run) | Not measured | 89.4% | Only GPT-5.2 measured |
| SciCode | 59.0% | Not measured | Only GLM 5.3 measured |
| Terminal-Bench Hard (Artificial Analysis run) | Not measured | 47.0% | Only GPT-5.2 measured |
| Terminal-Bench | Not measured | 64.9% | Only GPT-5.2 measured |
| FrontierCode | 40.1% | Not measured | Only GLM 5.3 measured |
| FrontierSWE | 30.2% | Not measured | Only GLM 5.3 measured |
| GSO-Bench | Not measured | 27.5% | Only GPT-5.2 measured |
| WeirdML | 75.4% | 72.2% | GLM 5.3 |
| ALE-Bench | 1317.4 | 1293.5 | GLM 5.3 |
| LMArena Text: coding | 1497 | 1442 | GLM 5.3 |
| LMArena Text: software and IT | 1493 | 1434 | GLM 5.3 |
| LMArena WebDev: React | 1641 | Not measured | Only GLM 5.3 measured |
| LMArena WebDev: HTML | 1595 | 1427 | GLM 5.3 |
| Neither model has a result on the other 10 boards here | |||
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index20
| Benchmark | GLM 5.3 | GPT-5.2 | Leader |
|---|---|---|---|
| AA Intelligence Index | 44.8 | 30.4 | GLM 5.3 |
| Epoch Capabilities Index | 156.0 | 153.7 | GLM 5.3 |
Agentic workLike-for-likePublic boards, shown as published, kept out of the index20
| Benchmark | GLM 5.3 | GPT-5.2 | Leader |
|---|---|---|---|
| AA Agentic Index | 53.1 | Not measured | Only GLM 5.3 measured |
| Design Arena: agents building web apps | Not measured | 1086 | Only GPT-5.2 measured |
| Design Arena: agents building full-stack apps | Not measured | 1043 | Only GPT-5.2 measured |
| Design Arena: agents building mobile apps | 1194 | 1093 | GLM 5.3 |
| Design Arena: agents building Android apps | Not measured | 1059 | Only GPT-5.2 measured |
| Design Arena: agents building Godot games | Not measured | 1142 | Only GPT-5.2 measured |
| Design Arena: HTML slides | 1189 | Not measured | Only GLM 5.3 measured |
| Design Arena: PowerPoint slides | 1230 | Not measured | Only GLM 5.3 measured |
| LiveBench research | 73.2% | Not measured | Only GLM 5.3 measured |
| LiveBench browser | 0.0% | Not measured | Only GLM 5.3 measured |
| LiveBench long agentic | 35.9% | Not measured | Only GLM 5.3 measured |
| SREBench (Vals AI) | 9.2% | Not measured | Only GLM 5.3 measured |
| Finance Agent (Vals AI) | 55.8% | Not measured | Only GLM 5.3 measured |
| Tax Agent Bench (Vals AI) | 39.7% | Not measured | Only GLM 5.3 measured |
| Legal Research Bench (Vals AI) | 49.0% | Not measured | Only GLM 5.3 measured |
| Harvey's Legal Agent Benchmark (Vals AI) | 14.6% | Not measured | Only GLM 5.3 measured |
| Excel Modeling Benchmark (Vals AI) | 56.3% | Not measured | Only GLM 5.3 measured |
| Terminal-Bench-Science (Vals AI) | 5.7% | Not measured | Only GLM 5.3 measured |
| τ²-Bench Telecom | Not measured | 84.8% | Only GPT-5.2 measured |
| APEX-Agents | 56.6% | Not measured | Only GLM 5.3 measured |
| METR time horizons | Not measured | 75.3% | Only GPT-5.2 measured |
| Remote Labor Index | Not measured | 2.5% | Only GPT-5.2 measured |
| GDPval | Not measured | 49.7% | Only GPT-5.2 measured |
| DeepResearch Bench | Not measured | 41.1% | Only GPT-5.2 measured |
| Vending-Bench 2 | $8,164 | $3,591 | GLM 5.3 |
| Berkeley Function Calling Leaderboard (BFCL) | Not measured | 55.9% | Only GPT-5.2 measured |
| Neither model has a result on the other 15 boards here | |||
ReasoningLike-for-likePublic boards, shown as published, kept out of the index63
| Benchmark | GLM 5.3 | GPT-5.2 | Leader |
|---|---|---|---|
| GPQA Diamond | 88.1% | 91.7% | GPT-5.2 |
| LiveBench reasoning | 85.8% | Not measured | Only GLM 5.3 measured |
| MysteryMechanism (Vals AI) | 23.0% | Not measured | Only GLM 5.3 measured |
| MMMU Pro (Vals AI) | Not measured | 86.7% | Only GPT-5.2 measured |
| GPQA Diamond (Artificial Analysis run) | 91.7% | 90.3% | GLM 5.3 |
| Humanity's Last Exam (Artificial Analysis run) | 42.3% | 37.7% | GLM 5.3 |
| GPQA Diamond (Epoch AI run) | 90.9% | 91.4% | GPT-5.2 |
| ARC-AGI | Not measured | 86.2% | Only GPT-5.2 measured |
| ARC-AGI-2 | Not measured | 52.9% | Only GPT-5.2 measured |
| SimpleBench | Not measured | 45.8% | Only GPT-5.2 measured |
| Chess puzzles | 37.0% | 48.0% | GPT-5.2 |
| Mystery game puzzles | 33.0% | 23.0% | GLM 5.3 |
| EBR-bench | Not measured | 23.0% | Only GPT-5.2 measured |
| VPCT | Not measured | 84.0% | Only GPT-5.2 measured |
| LMArena Text: hard prompts | 1490 | 1428 | GLM 5.3 |
| LMArena Text: expert | 1514 | 1445 | GLM 5.3 |
| LMArena Text: hard prompts in English | 1496 | 1431 | GLM 5.3 |
| LMArena Vision: diagrams | Not measured | 1265 | Only GPT-5.2 measured |
| LMArena Vision: homework | Not measured | 1276 | Only GPT-5.2 measured |
| Kagi LLM Benchmark | Not measured | 73.3% | Only GPT-5.2 measured |
| Neither model has a result on the other 2 boards here | |||
MathLike-for-likePublic boards, shown as published, kept out of the index42
| Benchmark | GLM 5.3 | GPT-5.2 | Leader |
|---|---|---|---|
| AIME | Not measured | 96.9% | Only GPT-5.2 measured |
| MGSM (Vals AI) | Not measured | 94.0% | Only GPT-5.2 measured |
| AA Math Index | Not measured | 99.0 | Only GPT-5.2 measured |
| AIME 2025 (Artificial Analysis run) | Not measured | 99.0% | Only GPT-5.2 measured |
| OTIS Mock AIME | 91.1% | 96.1% | GPT-5.2 |
| FrontierMath, tiers 1 to 3 | 68.8% | 67.4% | GLM 5.3 |
| FrontierMath, tier 4 | 29.3% | 31.7% | GPT-5.2 |
| ProofBench | 49.0% | 15.0% | GLM 5.3 |
| LMArena Text: math | 1495 | 1440 | GLM 5.3 |
| LMArena Text: mathematical work | 1497 | 1442 | GLM 5.3 |
| Neither model has a result on the other 3 boards here | |||
KnowledgeLike-for-likePublic boards, shown as published, kept out of the index72
| Benchmark | GLM 5.3 | GPT-5.2 | Leader |
|---|---|---|---|
| MMLU Pro | 86.8% | 86.2% | GLM 5.3 |
| LegalBench (Vals AI) | 84.8% | 82.8% | GLM 5.3 |
| CaseLaw (Vals AI) | Not measured | 66.0% | Only GPT-5.2 measured |
| CorpFin (Vals AI) | Not measured | 65.9% | Only GPT-5.2 measured |
| TaxEval (Vals AI) | 72.4% | 75.8% | GPT-5.2 |
| MortgageTax (Vals AI) | Not measured | 67.1% | Only GPT-5.2 measured |
| MedQA (Vals AI) | Not measured | 94.1% | Only GPT-5.2 measured |
| MedCode (Vals AI) | 42.9% | 49.8% | GPT-5.2 |
| MMLU Pro (Artificial Analysis run) | Not measured | 87.4% | Only GPT-5.2 measured |
| SimpleQA Verified | 41.0% | 37.1% | GLM 5.3 |
| LMArena Text: science | 1498 | 1425 | GLM 5.3 |
| LMArena Text: medicine and healthcare | 1481 | 1425 | GLM 5.3 |
| LMArena Text: legal and government | 1473 | 1427 | GLM 5.3 |
| LMArena Text: business and finance | 1459 | 1414 | GLM 5.3 |
| FACTS Benchmark Suite | Not measured | 48.5% | Only GPT-5.2 measured |
| Neither model has a result on the other 3 boards here | |||
Web designLike-for-likePublic boards, shown as published, kept out of the index80
| Benchmark | GLM 5.3 | GPT-5.2 | Leader |
|---|---|---|---|
| Design Arena: websites | 1295 | 1195 | GLM 5.3 |
| Design Arena: UI components | 1314 | 1190 | GLM 5.3 |
| Design Arena: game dev | 1311 | 1195 | GLM 5.3 |
| Design Arena: data viz | 1260 | 1202 | GLM 5.3 |
| Design Arena: 3D | 1332 | 1079 | GLM 5.3 |
| Design Arena: all code categories | 1304 | 1170 | GLM 5.3 |
| Design Arena: SVG | 1294 | 1139 | GLM 5.3 |
| Design Arena: ASCII art | 1208 | 1152 | GLM 5.3 |
| LiveBench design | 77.3% | Not measured | Only GLM 5.3 measured |
GeneralLike-for-likePublic boards, shown as published, kept out of the index192
| Benchmark | GLM 5.3 | GPT-5.2 | Leader |
|---|---|---|---|
| Vals Index | 59.7% | Not measured | Only GLM 5.3 measured |
| LiveBench data analysis | 70.2% | Not measured | Only GLM 5.3 measured |
| Public Benefits Bench (Vals AI) | 68.5% | Not measured | Only GLM 5.3 measured |
| SAGE (Vals AI) | Not measured | 49.3% | Only GPT-5.2 measured |
| MedScribe (Vals AI) | 88.8% | 84.4% | GLM 5.3 |
| IFBench | Not measured | 75.4% | Only GPT-5.2 measured |
| Long context reasoning (AA-LCR) | 79.7% | 82.7% | GPT-5.2 |
| DTBench | 87.7% | 90.9% | GPT-5.2 |
| LMCA | 55.5% | 43.9% | GLM 5.3 |
| CL-bench | Not measured | 18.2% | Only GPT-5.2 measured |
| Furniture assembly | Not measured | 38.3% | Only GPT-5.2 measured |
| LMArena Text | 1474 | 1416 | GLM 5.3 |
| LMArena Text: instruction following | 1479 | 1409 | GLM 5.3 |
| LMArena Text: creative writing | 1458 | 1376 | GLM 5.3 |
| LMArena Vision | Not measured | 1246 | Only GPT-5.2 measured |
| LMArena Text: English | 1482 | 1422 | GLM 5.3 |
| LMArena Text: non-English | 1459 | 1405 | GLM 5.3 |
| LMArena Text: Chinese | 1532 | 1447 | GLM 5.3 |
| LMArena Text: German | 1488 | 1444 | GLM 5.3 |
| LMArena Text: Japanese | 1460 | 1420 | GLM 5.3 |
| LMArena Text: Korean | 1460 | 1374 | GLM 5.3 |
| LMArena Text: French | 1503 | 1454 | GLM 5.3 |
| LMArena Text: Spanish | 1462 | 1413 | GLM 5.3 |
| LMArena Text: Russian | 1469 | 1415 | GLM 5.3 |
| LMArena Text: Polish | 1466 | 1426 | GLM 5.3 |
| LMArena Text: multi-turn | 1472 | 1419 | GLM 5.3 |
| LMArena Text: longer queries | 1483 | 1413 | GLM 5.3 |
| LMArena Text: writing and language | 1469 | 1392 | GLM 5.3 |
| LMArena Text: entertainment and media | 1447 | 1378 | GLM 5.3 |
| LMArena Vision: English | Not measured | 1243 | Only GPT-5.2 measured |
| LMArena Vision: Chinese | Not measured | 1305 | Only GPT-5.2 measured |
| LMArena Vision: reading text | Not measured | 1257 | Only GPT-5.2 measured |
| LMArena Vision: humor | Not measured | 1227 | Only GPT-5.2 measured |
| LMArena Vision: captioning | Not measured | 1267 | Only GPT-5.2 measured |
| LMArena Vision: recognizing things | Not measured | 1232 | Only GPT-5.2 measured |
| LMArena Vision: creative writing | Not measured | 1225 | Only GPT-5.2 measured |
| LMArena Document | Not measured | 1405 | Only GPT-5.2 measured |
| Neither model has a result on the other 9 boards here | |||
Image generationNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
| Benchmark | GLM 5.3 | GPT-5.2 | Leader |
|---|---|---|---|
| Neither model has a result on 5 boards here | |||
What each workload costs
Three fixed token mixes at each model's listed price, so per-token rates become a bill.
Chat turn
1K input + 500 output tokens
GLM 5.3 costs less
Repository review
50K input + 3K output tokens
GLM 5.3 costs less
Agent loop
200K input + 10K output tokens
GLM 5.3 costs less
Listed OpenRouter rates, without cache or batch discounts. A workload that does not fit a model's context window would need splitting.
Specification differences
The leading value in a row is marked; a row with a missing value names no leader. Our two rows are marked in red.
| Specification | Z.ai GLM 5.3 | OpenAI GPT-5.2 |
|---|---|---|
| Maker | Z.ai | OpenAI |
| Weights | Open weights | No public weights listed |
| Released | August 18, 2026 | December 10, 2025 |
| Context window | 1.05M | 400K |
| Input $/Mtok | $1.40 | $1.75 |
| Output $/Mtok | $4.40 | $14.00 |
| Coding index | 70.6 | 55.9 |
| Independent measurements behind it | 6 | 3 |
| Cost per solved task | $0.36 | $1.91 |
| Source | OpenRouter listing | OpenRouter listing |
Questions
Is GLM 5.3 better than GPT-5.2 for coding?
Pick GLM 5.3 for coding: it scores 70.6 on the Vibe Coding Index against 55.9 for GPT-5.2. Pick GPT-5.2 for work measured by LiveCodeBench (Vals AI) or GPQA Diamond, where it leads.
Which is cheaper, GLM 5.3 or GPT-5.2?
GLM 5.3 costs $1.40 per million input tokens and $4.40 per million output tokens; GPT-5.2 costs $1.75 and $14.00. Per solved SWE-bench Verified task, GLM 5.3 costs $0.36 and GPT-5.2 $1.91.
Which has the bigger context window, GLM 5.3 or GPT-5.2?
GLM 5.3, with 1.05M tokens against 400K.
How many benchmarks is the GLM 5.3 vs GPT-5.2 comparison based on?
66 benchmarks have a result for both models as of October 22, 2026. GLM 5.3's coding index rests on 6 independent benchmarks and GPT-5.2's on 3. Where only one model has a result, the page says so instead of naming a winner.