Model comparison
GLM-5.2 vs GPT-5.2
GPT-5.2 is the stronger model overall, scoring 54.1 to 51.1 on the Noometry Index. GLM-5.2 costs 2.2× less per token, which makes it the better buy when GPT-5.2's lead doesn't matter for your workload.
Last verified . 42 shared benchmarks.
Summary
- They share 42 benchmarks with published results for both. GLM-5.2 scores higher in 4 categories and GPT-5.2 in 5 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where GPT-5.2 leads 50.2 to 42.3.
- The biggest single-benchmark swing is ARC-AGI-2: 22.8% for GLM-5.2 and 52.9% for GPT-5.2.
- GLM-5.2 is cheaper at $1.40 / $4.40 per million input/output tokens, against $1.75 / $14 for GPT-5.2.
- GLM-5.2 accepts more context: 1M tokens versus 400K.
- GLM-5.2 has downloadable open weights; the other is API-only.
Side by side
| GLM-5.2 | GPT-5.2 | |
|---|---|---|
| Provider | Z.ai (Zhipu) | OpenAI |
| Noometry Index | 51.1 | 54.1 |
| Released | 2026-06-13 | 2025-12-11 |
| Weights | Open | Proprietary |
| Context window | 1M | 400K |
| Max output | 131K | 128K |
| Input $ / M tokens | $1.40 | $1.75 |
| Output $ / M tokens | $4.40 | $14 |
| Results tracked | 51 | 67 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Too close to call
GLM-5.2: 51.3 (#41), GPT-5.2: 51.6 (#37)
| Benchmark | GLM-5.2 | GPT-5.2 |
|---|---|---|
| SWE-bench Verified | 78.7% | 73.8% |
| LMArena WebDev | 1603 | 1416 |
| WeirdML | 70.1% | 72.2% |
| LMArena Coding | 1485 | 1447 |
| ALE-Bench | 1,047 | 1,294 |
| DeepSWE | 43.8% | — |
| FrontierCode | 24.5% | — |
| SWE-bench Verified (bash only) | — | 72.8% |
| SWE-bench Multilingual | — | 66.7% |
| SciCode | 50.5% | — |
| GSO | — | 27.4% |
| AlgoTune | — | 2.05 |
Agentic & Tool Use GPT-5.2 leads
GLM-5.2: 32.4 (#63), GPT-5.2: 40.2 (#24)
| Benchmark | GLM-5.2 | GPT-5.2 |
|---|---|---|
| τ²-bench Banking | 37.1% | 32.2% |
| Vending-Bench 2 | 8,314 | 3,591 |
| Terminal-Bench | — | 64.9% |
| APEX-Agents | 45.2% | — |
| Berkeley Function Calling Leaderboard | — | 55.9% |
| GDPval | — | 49.7% |
| Remote Labor Index | — | 2.5% |
| τ²-bench Airline | — | 83% |
| τ²-bench Retail | — | 81.6% |
| τ²-bench Telecom | — | 89.7% |
| DeepResearch Bench | — | 41.1% |
| PostTrainBench | 31.7% | — |
| GBAEval | 0% | — |
| LMArena Search | — | 1207 |
| METR Time Horizons | — | 75.3% |
Reasoning GPT-5.2 leads
GLM-5.2: 42.3 (#52), GPT-5.2: 50.2 (#35)
| Benchmark | GLM-5.2 | GPT-5.2 |
|---|---|---|
| ARC-AGI-2 | 22.8% | 52.9% |
| SimpleBench | 58.8% | 45.8% |
| Kagi LLM Benchmark | 62.6% | 73.3% |
| NYT Connections (extended) | 74.3% | 83.6% |
| ARC-AGI-1 | 77% | 86.2% |
| Chess Puzzles | 21% | 49% |
| EBR-Bench | 9.5% | 23% |
| LMArena Hard Prompts | 1480 | 1445 |
| Mystery Game Puzzles | 19% | 23% |
| DTBench | 93.6% | 90.9% |
| LMCA | 45.8% | 43.9% |
| Epoch Capabilities Index | 151.78 | 153.45 |
| CritPt | 20.9% | — |
| EnigmaEval | — | 10.4% |
| Surface Evolver Bench | 55.6% | — |
| ForecastBench | — | 60.1 |
Math GPT-5.2 leads
GLM-5.2: 55.7 (#43), GPT-5.2: 60.0 (#38)
| Benchmark | GLM-5.2 | GPT-5.2 |
|---|---|---|
| FrontierMath (Tiers 1-3) | 59.2% | 67.4% |
| FrontierMath Tier 4 | 29.3% | 31.7% |
| MathArena Final-Answer Competitions | 67.6% | 72% |
| OTIS Mock AIME 2024-2025 | 86.4% | 96.1% |
| ProofBench | 35% | 15% |
| LMArena Math | 1482 | 1440 |
| FrontierMath (Feb 2025 set) | — | 40.7% |
| FrontierMath Tier 4 (v1) | — | 18.8% |
Knowledge GPT-5.2 leads
GLM-5.2: 57.1 (#40), GPT-5.2: 59.3 (#32)
| Benchmark | GLM-5.2 | GPT-5.2 |
|---|---|---|
| GPQA Diamond | 91.9% | 91.4% |
| SimpleQA Verified | 34.2% | 37.1% |
| LMArena Expert | 1486 | 1445 |
| Humanity's Last Exam | — | 27.8% |
| Vectara Hallucination Rate | — | 8.4% |
Multimodal Not comparable
GLM-5.2: —, GPT-5.2: 51.3 (#7)
| Benchmark | GLM-5.2 | GPT-5.2 |
|---|---|---|
| LMArena Vision | — | 1268 |
| VPCT | — | 84% |
| Furniture Assembly | — | 38.3% |
| LMArena Document | — | 1405 |
Multilingual GLM-5.2 leads
GLM-5.2: 55.8 (#26), GPT-5.2: 53.4 (#67)
| Benchmark | GLM-5.2 | GPT-5.2 |
|---|---|---|
| LMArena Non-English | 1459 | 1425 |
| LMArena Chinese | 1519 | 1460 |
| LMArena French | 1479 | 1455 |
| LMArena German | 1468 | 1448 |
| LMArena Japanese | 1451 | 1420 |
| LMArena Korean | 1445 | 1392 |
| LMArena Russian | 1466 | 1440 |
| LMArena Spanish | 1477 | 1433 |
Instruction Following GLM-5.2 leads
GLM-5.2: 76.9 (#34), GPT-5.2: 74.7 (#89)
| Benchmark | GLM-5.2 | GPT-5.2 |
|---|---|---|
| LMArena Instruction Following | 1465 | 1417 |
Long Context GLM-5.2 leads
GLM-5.2: 45.3 (#43), GPT-5.2: 44.0 (#78)
| Benchmark | GLM-5.2 | GPT-5.2 |
|---|---|---|
| LMArena Longer Query | 1479 | 1428 |
| CL-bench | — | 18.2% |
Writing & Preference GLM-5.2 leads
GLM-5.2: 70.4 (#21), GPT-5.2: 66.8 (#32)
| Benchmark | GLM-5.2 | GPT-5.2 |
|---|---|---|
| LMArena Text | 1470 | 1439 |
| LMArena Creative Writing | 1462 | 1401 |
| EQ-Bench Creative Writing | 1757 | 1703 |
| LMArena Multi-Turn | 1469 | 1458 |
| EQ-Bench 4 | 1222 | — |
Frequently asked questions
Is GLM-5.2 better than GPT-5.2?
GPT-5.2 is the stronger model overall, scoring 54.1 to 51.1 on the Noometry Index. GLM-5.2 costs 2.2× less per token, which makes it the better buy when GPT-5.2's lead doesn't matter for your workload.
Which is cheaper, GLM-5.2 or GPT-5.2?
GLM-5.2 is cheaper. It lists at $1.40 per million input tokens and $4.40 per million output tokens; GPT-5.2 lists at $1.75 and $14.
Is GLM-5.2 or GPT-5.2 better for coding?
They score almost the same on coding (51.3 vs 51.6); test both on your own repository before choosing.
Which has the bigger context window?
GLM-5.2 does, with 1M tokens against 400K.
How many benchmarks do GLM-5.2 and GPT-5.2 share?
42 benchmarks have published results for both models. GLM-5.2 has 51 scored results on Noometry and GPT-5.2 has 67.