Model comparison
GLM-5.2 vs o3
GLM-5.2 is the stronger model overall, scoring 51.1 to 47.5 on the Noometry Index.
Last verified . 35 shared benchmarks.
Summary
- They share 35 benchmarks with published results for both. GLM-5.2 scores higher in 7 categories and o3 in 2 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where GLM-5.2 leads 42.3 to 32.0.
- The biggest single-benchmark swing is FrontierMath (Tiers 1-3): 59.2% for GLM-5.2 and 33.3% for o3.
- GLM-5.2 is cheaper at $1.40 / $4.40 per million input/output tokens, against $2 / $8 for o3.
- GLM-5.2 accepts more context: 1M tokens versus 200K.
- GLM-5.2 has downloadable open weights; the other is API-only.
Side by side
| GLM-5.2 | o3 | |
|---|---|---|
| Provider | Z.ai (Zhipu) | OpenAI |
| Noometry Index | 51.1 | 47.5 |
| Released | 2026-06-13 | 2025-04-16 |
| Weights | Open | Proprietary |
| Context window | 1M | 200K |
| Max output | 131K | 100K |
| Input $ / M tokens | $1.40 | $2 |
| Output $ / M tokens | $4.40 | $8 |
| Results tracked | 51 | 63 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding GLM-5.2 leads
GLM-5.2: 51.3 (#41), o3: 46.8 (#64)
| Benchmark | GLM-5.2 | o3 |
|---|---|---|
| SWE-bench Verified | 78.7% | 62.3% |
| WeirdML | 70.1% | 52.4% |
| LMArena Coding | 1485 | 1408 |
| ALE-Bench | 1,047 | 933.55 |
| DeepSWE | 43.8% | — |
| FrontierCode | 24.5% | — |
| SWE-bench Verified (bash only) | — | 58.4% |
| Aider Polyglot | — | 81.3% |
| LMArena WebDev | 1603 | — |
| SciCode | 50.5% | — |
| GSO | — | 8.8% |
| CadEval | — | 74% |
Agentic & Tool Use o3 leads
GLM-5.2: 32.4 (#63), o3: 34.5 (#44)
| Benchmark | GLM-5.2 | o3 |
|---|---|---|
| APEX-Agents | 45.2% | — |
| Berkeley Function Calling Leaderboard | — | 63% |
| GDPval | — | 30.8% |
| τ²-bench Banking | 37.1% | — |
| DeepResearch Bench | — | 45.2% |
| OSWorld | — | 23% |
| PostTrainBench | 31.7% | — |
| GBAEval | 0% | — |
| LMArena Search | — | 1144 |
| METR Time Horizons | — | 65.4% |
| Vending-Bench 2 | 8,314 | — |
Reasoning GLM-5.2 leads
GLM-5.2: 42.3 (#52), o3: 32.0 (#78)
| Benchmark | GLM-5.2 | o3 |
|---|---|---|
| ARC-AGI-2 | 22.8% | 6.5% |
| SimpleBench | 58.8% | 53.1% |
| Kagi LLM Benchmark | 62.6% | 67.6% |
| ARC-AGI-1 | 77% | 60.8% |
| CritPt | 20.9% | 1.4% |
| Chess Puzzles | 21% | 38% |
| LMArena Hard Prompts | 1480 | 1402 |
| Mystery Game Puzzles | 19% | 29% |
| DTBench | 93.6% | 84.8% |
| LMCA | 45.8% | 39.7% |
| Epoch Capabilities Index | 151.78 | 146.86 |
| NYT Connections (extended) | 74.3% | — |
| EnigmaEval | — | 13.1% |
| EBR-Bench | 9.5% | — |
| Surface Evolver Bench | 55.6% | — |
| ForecastBench | — | 62.5 |
Math GLM-5.2 leads
GLM-5.2: 55.7 (#43), o3: 50.2 (#58)
| Benchmark | GLM-5.2 | o3 |
|---|---|---|
| FrontierMath (Tiers 1-3) | 59.2% | 33.3% |
| OTIS Mock AIME 2024-2025 | 86.4% | 84.4% |
| LMArena Math | 1482 | 1426 |
| FrontierMath Tier 4 | 29.3% | — |
| MathArena Final-Answer Competitions | 67.6% | — |
| ProofBench | 35% | — |
| Omni-MATH | — | 71.4% |
| MATH Level 5 | — | 97.8% |
| FrontierMath (Feb 2025 set) | — | 18.7% |
| FrontierMath Tier 4 (v1) | — | 2.1% |
Knowledge GLM-5.2 leads
GLM-5.2: 57.1 (#40), o3: 54.6 (#52)
| Benchmark | GLM-5.2 | o3 |
|---|---|---|
| GPQA Diamond | 91.9% | 81.8% |
| SimpleQA Verified | 34.2% | 49.4% |
| LMArena Expert | 1486 | 1402 |
| Humanity's Last Exam | — | 20.3% |
| MMLU-Pro | — | 85.9% |
| Confabulations | — | 14.4% |
| GPQA (HELM) | — | 75.3% |
Multimodal Not comparable
GLM-5.2: —, o3: 41.4 (#36)
| Benchmark | GLM-5.2 | o3 |
|---|---|---|
| LMArena Vision | — | 1214 |
| GeoBench | — | 74% |
| VPCT | — | 52% |
Multilingual GLM-5.2 leads
GLM-5.2: 55.8 (#26), o3: 51.7 (#105)
| Benchmark | GLM-5.2 | o3 |
|---|---|---|
| LMArena Non-English | 1459 | 1401 |
| LMArena Chinese | 1519 | 1437 |
| LMArena French | 1479 | 1430 |
| LMArena German | 1468 | 1420 |
| LMArena Japanese | 1451 | 1403 |
| LMArena Korean | 1445 | 1370 |
| LMArena Russian | 1466 | 1406 |
| LMArena Spanish | 1477 | 1395 |
Instruction Following GLM-5.2 leads
GLM-5.2: 76.9 (#34), o3: 72.8 (#127)
| Benchmark | GLM-5.2 | o3 |
|---|---|---|
| LMArena Instruction Following | 1465 | 1368 |
| IFEval | — | 86.9% |
Long Context o3 leads
GLM-5.2: 45.3 (#43), o3: 53.3 (#6)
| Benchmark | GLM-5.2 | o3 |
|---|---|---|
| LMArena Longer Query | 1479 | 1372 |
| Fiction.LiveBench | — | 88.9% |
| CL-bench | — | 17.8% |
Writing & Preference GLM-5.2 leads
GLM-5.2: 70.4 (#21), o3: 63.5 (#64)
| Benchmark | GLM-5.2 | o3 |
|---|---|---|
| LMArena Text | 1470 | 1410 |
| LMArena Creative Writing | 1462 | 1359 |
| EQ-Bench Creative Writing | 1757 | 1676 |
| LMArena Multi-Turn | 1469 | 1405 |
| Short-Story Creative Writing | — | 83.9% |
| WildBench | — | 86.1% |
| EQ-Bench 4 | 1222 | — |
Frequently asked questions
Is GLM-5.2 better than o3?
GLM-5.2 is the stronger model overall, scoring 51.1 to 47.5 on the Noometry Index.
Which is cheaper, GLM-5.2 or o3?
GLM-5.2 is cheaper. It lists at $1.40 per million input tokens and $4.40 per million output tokens; o3 lists at $2 and $8.
Is GLM-5.2 or o3 better for coding?
GLM-5.2 scores higher on coding benchmarks: 51.3 versus 46.8 in the Noometry coding category.
Which has the bigger context window?
GLM-5.2 does, with 1M tokens against 200K.
How many benchmarks do GLM-5.2 and o3 share?
35 benchmarks have published results for both models. GLM-5.2 has 51 scored results on Noometry and o3 has 63.