Model comparison
GLM-5.2 vs GPT-5.6 Sol
GPT-5.6 Sol is the stronger model overall, scoring 65.0 to 51.1 on the Noometry Index. GLM-5.2 costs 3.7× less per token, which makes it the better buy when GPT-5.6 Sol's lead doesn't matter for your workload.
Last verified . 49 shared benchmarks.
Summary
- They share 49 benchmarks with published results for both. GLM-5.2 scores higher in 1 category and GPT-5.6 Sol in 8 categories; 6 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where GPT-5.6 Sol leads 74.8 to 42.3.
- The biggest single-benchmark swing is ARC-AGI-2: 22.8% for GLM-5.2 and 92.5% for GPT-5.6 Sol.
- GLM-5.2 is cheaper at $1.40 / $4.40 per million input/output tokens, against $4 / $20 for GPT-5.6 Sol.
- GPT-5.6 Sol accepts more context: 1.05M tokens versus 1M.
- GLM-5.2 has downloadable open weights; the other is API-only.
Side by side
| GLM-5.2 | GPT-5.6 Sol | |
|---|---|---|
| Provider | Z.ai (Zhipu) | OpenAI |
| Noometry Index | 51.1 | 65.0 |
| Released | 2026-06-13 | 2026-07-09 |
| Weights | Open | Proprietary |
| Context window | 1M | 1.05M |
| Max output | 131K | 128K |
| Input $ / M tokens | $1.40 | $4 |
| Output $ / M tokens | $4.40 | $20 |
| Results tracked | 51 | 65 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding GPT-5.6 Sol leads
GLM-5.2: 51.3 (#41), GPT-5.6 Sol: 65.1 (#7)
| Benchmark | GLM-5.2 | GPT-5.6 Sol |
|---|---|---|
| DeepSWE | 43.8% | 72.7% |
| FrontierCode | 24.5% | 47.5% |
| LMArena WebDev | 1603 | 1618 |
| SciCode | 50.5% | 57.1% |
| WeirdML | 70.1% | 89.4% |
| LMArena Coding | 1485 | 1498 |
| ALE-Bench | 1,047 | 2,177 |
| SWE-bench Verified | 78.7% | — |
| CursorBench | — | 41.7% |
| FrontierSWE | — | 32.2% |
| GSO | — | 76.5% |
| MirrorCode | — | 20% |
Agentic & Tool Use GPT-5.6 Sol leads
GLM-5.2: 32.4 (#63), GPT-5.6 Sol: 50.3 (#7)
| Benchmark | GLM-5.2 | GPT-5.6 Sol |
|---|---|---|
| APEX-Agents | 45.2% | 51.4% |
| τ²-bench Banking | 37.1% | 46.9% |
| PostTrainBench | 31.7% | 36.2% |
| GBAEval | 0% | 52.6% |
| Vending-Bench 2 | 8,314 | 9,619 |
| OSWorld 2.0 | — | 27.3% |
| BALROG | — | 60% |
| GDP.pdf | — | 30.7% |
| LMArena Search | — | 1257 |
Reasoning GPT-5.6 Sol leads
GLM-5.2: 42.3 (#52), GPT-5.6 Sol: 74.8 (#8)
| Benchmark | GLM-5.2 | GPT-5.6 Sol |
|---|---|---|
| ARC-AGI-2 | 22.8% | 92.5% |
| SimpleBench | 58.8% | 71.7% |
| Kagi LLM Benchmark | 62.6% | 67% |
| NYT Connections (extended) | 74.3% | 93.8% |
| ARC-AGI-1 | 77% | 97.5% |
| CritPt | 20.9% | 32.3% |
| Chess Puzzles | 21% | 64% |
| EBR-Bench | 9.5% | 44.8% |
| LMArena Hard Prompts | 1480 | 1484 |
| Mystery Game Puzzles | 19% | 58% |
| DTBench | 93.6% | 96% |
| LMCA | 45.8% | 59.2% |
| Surface Evolver Bench | 55.6% | 93.1% |
| Epoch Capabilities Index | 151.78 | 161.66 |
| EnigmaEval | — | 37.1% |
| Bench to the Future 3 | — | 0.14 |
Math GPT-5.6 Sol leads
GLM-5.2: 55.7 (#43), GPT-5.6 Sol: 85.6 (#9)
| Benchmark | GLM-5.2 | GPT-5.6 Sol |
|---|---|---|
| FrontierMath (Tiers 1-3) | 59.2% | 89.1% |
| FrontierMath Tier 4 | 29.3% | 82.9% |
| OTIS Mock AIME 2024-2025 | 86.4% | 100% |
| ProofBench | 35% | 83% |
| LMArena Math | 1482 | 1474 |
| MathArena Final-Answer Competitions | 67.6% | — |
| FrontierMath Erdős | — | 0% |
Knowledge GPT-5.6 Sol leads
GLM-5.2: 57.1 (#40), GPT-5.6 Sol: 64.3 (#18)
| Benchmark | GLM-5.2 | GPT-5.6 Sol |
|---|---|---|
| GPQA Diamond | 91.9% | 93.5% |
| SimpleQA Verified | 34.2% | 69.7% |
| LMArena Expert | 1486 | 1516 |
| Vectara Hallucination Rate | — | 12.4% |
Multimodal Not comparable
GLM-5.2: —, GPT-5.6 Sol: 48.6 (#9)
| Benchmark | GLM-5.2 | GPT-5.6 Sol |
|---|---|---|
| LMArena Vision | — | 1281 |
| Blueprint-Bench 2 | — | 33.6% |
| Furniture Assembly | — | 56.7% |
| LMArena Document | — | 1483 |
Multilingual Too close to call
GLM-5.2: 55.8 (#26), GPT-5.6 Sol: 55.3 (#32)
| Benchmark | GLM-5.2 | GPT-5.6 Sol |
|---|---|---|
| LMArena Non-English | 1459 | 1452 |
| LMArena Chinese | 1519 | 1527 |
| LMArena French | 1479 | 1477 |
| LMArena German | 1468 | 1476 |
| LMArena Japanese | 1451 | 1471 |
| LMArena Korean | 1445 | 1442 |
| LMArena Russian | 1466 | 1468 |
| LMArena Spanish | 1477 | 1441 |
Instruction Following Too close to call
GLM-5.2: 76.9 (#34), GPT-5.6 Sol: 77.7 (#16)
| Benchmark | GLM-5.2 | GPT-5.6 Sol |
|---|---|---|
| LMArena Instruction Following | 1465 | 1482 |
Long Context Too close to call
GLM-5.2: 45.3 (#43), GPT-5.6 Sol: 45.4 (#42)
| Benchmark | GLM-5.2 | GPT-5.6 Sol |
|---|---|---|
| LMArena Longer Query | 1479 | 1480 |
Writing & Preference GPT-5.6 Sol leads
GLM-5.2: 70.4 (#21), GPT-5.6 Sol: 73.3 (#12)
| Benchmark | GLM-5.2 | GPT-5.6 Sol |
|---|---|---|
| LMArena Text | 1470 | 1457 |
| LMArena Creative Writing | 1462 | 1448 |
| EQ-Bench Creative Writing | 1757 | 1972 |
| EQ-Bench 4 | 1222 | 1250 |
| LMArena Multi-Turn | 1469 | 1460 |
Frequently asked questions
Is GLM-5.2 better than GPT-5.6 Sol?
GPT-5.6 Sol is the stronger model overall, scoring 65.0 to 51.1 on the Noometry Index. GLM-5.2 costs 3.7× less per token, which makes it the better buy when GPT-5.6 Sol's lead doesn't matter for your workload.
Which is cheaper, GLM-5.2 or GPT-5.6 Sol?
GLM-5.2 is cheaper. It lists at $1.40 per million input tokens and $4.40 per million output tokens; GPT-5.6 Sol lists at $4 and $20.
Is GLM-5.2 or GPT-5.6 Sol better for coding?
GPT-5.6 Sol scores higher on coding benchmarks: 65.1 versus 51.3 in the Noometry coding category.
Which has the bigger context window?
GPT-5.6 Sol does, with 1.05M tokens against 1M.
How many benchmarks do GLM-5.2 and GPT-5.6 Sol share?
49 benchmarks have published results for both models. GLM-5.2 has 51 scored results on Noometry and GPT-5.6 Sol has 65.