Model comparison
GLM-5.2 vs GLM-5.3-Flash
GLM-5.2 and GLM-5.3-Flash score almost the same on the Noometry Index (51.1 vs 51.8), so choose on price, context window or the category you care about most.
Last verified . 35 shared benchmarks.
Summary
- They share 35 benchmarks with published results for both. GLM-5.2 scores higher in 2 categories and GLM-5.3-Flash in 7 categories; 6 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where GLM-5.3-Flash leads 48.0 to 42.3.
- The biggest single-benchmark swing is ARC-AGI-2: 22.8% for GLM-5.2 and 65.8% for GLM-5.3-Flash.
- GLM-5.3-Flash is cheaper at $0.15 / $0.50 per million input/output tokens, against $1.40 / $4.40 for GLM-5.2.
Side by side
| GLM-5.2 | GLM-5.3-Flash | |
|---|---|---|
| Provider | Z.ai (Zhipu) | Z.ai (Zhipu) |
| Noometry Index | 51.1 | 51.8 |
| Released | 2026-06-13 | 2026-08-20 |
| Weights | Open | Open |
| Context window | 1M | 1M |
| Max output | 131K | 131K |
| Input $ / M tokens | $1.40 | $0.15 |
| Output $ / M tokens | $4.40 | $0.50 |
| Results tracked | 51 | 40 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding GLM-5.3-Flash leads
GLM-5.2: 51.3 (#41), GLM-5.3-Flash: 53.1 (#31)
| Benchmark | GLM-5.2 | GLM-5.3-Flash |
|---|---|---|
| DeepSWE | 43.8% | 63.4% |
| FrontierCode | 24.5% | 31.8% |
| LMArena WebDev | 1603 | 1609 |
| SciCode | 50.5% | 51.6% |
| LMArena Coding | 1485 | 1508 |
| ALE-Bench | 1,047 | 303.55 |
| SWE-bench Verified | 78.7% | — |
| CursorBench | — | 36.8% |
| FrontierSWE | — | 18.1% |
| WeirdML | 70.1% | — |
Agentic & Tool Use GLM-5.3-Flash leads
GLM-5.2: 32.4 (#63), GLM-5.3-Flash: 34.2 (#47)
| Benchmark | GLM-5.2 | GLM-5.3-Flash |
|---|---|---|
| APEX-Agents | 45.2% | 52.8% |
| τ²-bench Banking | 37.1% | — |
| PostTrainBench | 31.7% | — |
| GBAEval | 0% | — |
| GDP.pdf | — | 14% |
| Vending-Bench 2 | 8,314 | — |
Reasoning GLM-5.3-Flash leads
GLM-5.2: 42.3 (#52), GLM-5.3-Flash: 48.0 (#42)
| Benchmark | GLM-5.2 | GLM-5.3-Flash |
|---|---|---|
| ARC-AGI-2 | 22.8% | 65.8% |
| ARC-AGI-1 | 77% | 91% |
| CritPt | 20.9% | 15.4% |
| Chess Puzzles | 21% | 14% |
| LMArena Hard Prompts | 1480 | 1491 |
| Mystery Game Puzzles | 19% | 8% |
| Surface Evolver Bench | 55.6% | 52.5% |
| Epoch Capabilities Index | 151.78 | 151.88 |
| SimpleBench | 58.8% | — |
| Kagi LLM Benchmark | 62.6% | — |
| NYT Connections (extended) | 74.3% | — |
| EBR-Bench | 9.5% | — |
| DTBench | 93.6% | — |
| LMCA | 45.8% | — |
| Bench to the Future 3 | — | 0.15 |
Math GLM-5.2 leads
GLM-5.2: 55.7 (#43), GLM-5.3-Flash: 53.3 (#47)
| Benchmark | GLM-5.2 | GLM-5.3-Flash |
|---|---|---|
| FrontierMath (Tiers 1-3) | 59.2% | 55.8% |
| FrontierMath Tier 4 | 29.3% | 17.1% |
| OTIS Mock AIME 2024-2025 | 86.4% | 93.9% |
| ProofBench | 35% | 21% |
| LMArena Math | 1482 | 1500 |
| MathArena Final-Answer Competitions | 67.6% | — |
Knowledge GLM-5.3-Flash leads
GLM-5.2: 57.1 (#40), GLM-5.3-Flash: 58.4 (#36)
| Benchmark | GLM-5.2 | GLM-5.3-Flash |
|---|---|---|
| GPQA Diamond | 91.9% | 90.2% |
| LMArena Expert | 1486 | 1513 |
| SimpleQA Verified | 34.2% | — |
Multimodal Not comparable
GLM-5.2: —, GLM-5.3-Flash: 42.8 (#27)
| Benchmark | GLM-5.2 | GLM-5.3-Flash |
|---|---|---|
| LMArena Vision | — | 1296 |
Multilingual Too close to call
GLM-5.2: 55.8 (#26), GLM-5.3-Flash: 56.0 (#25)
| Benchmark | GLM-5.2 | GLM-5.3-Flash |
|---|---|---|
| LMArena Non-English | 1459 | 1462 |
| LMArena Chinese | 1519 | 1527 |
| LMArena French | 1479 | 1496 |
| LMArena German | 1468 | 1470 |
| LMArena Japanese | 1451 | 1429 |
| LMArena Korean | 1445 | 1446 |
| LMArena Russian | 1466 | 1469 |
| LMArena Spanish | 1477 | 1471 |
Instruction Following Too close to call
GLM-5.2: 76.9 (#34), GLM-5.3-Flash: 77.5 (#20)
| Benchmark | GLM-5.2 | GLM-5.3-Flash |
|---|---|---|
| LMArena Instruction Following | 1465 | 1478 |
Long Context Too close to call
GLM-5.2: 45.3 (#43), GLM-5.3-Flash: 45.4 (#39)
| Benchmark | GLM-5.2 | GLM-5.3-Flash |
|---|---|---|
| LMArena Longer Query | 1479 | 1482 |
Writing & Preference GLM-5.2 leads
GLM-5.2: 70.4 (#21), GLM-5.3-Flash: 65.3 (#50)
| Benchmark | GLM-5.2 | GLM-5.3-Flash |
|---|---|---|
| LMArena Text | 1470 | 1471 |
| LMArena Creative Writing | 1462 | 1442 |
| LMArena Multi-Turn | 1469 | 1467 |
| EQ-Bench Creative Writing | 1757 | — |
| EQ-Bench 4 | 1222 | — |
Frequently asked questions
Is GLM-5.2 better than GLM-5.3-Flash?
GLM-5.2 and GLM-5.3-Flash score almost the same on the Noometry Index (51.1 vs 51.8), so choose on price, context window or the category you care about most.
Which is cheaper, GLM-5.2 or GLM-5.3-Flash?
GLM-5.3-Flash is cheaper. It lists at $0.15 per million input tokens and $0.50 per million output tokens; GLM-5.2 lists at $1.40 and $4.40.
Is GLM-5.2 or GLM-5.3-Flash better for coding?
GLM-5.3-Flash scores higher on coding benchmarks: 53.1 versus 51.3 in the Noometry coding category.
Which has the bigger context window?
Both accept 1M tokens.
How many benchmarks do GLM-5.2 and GLM-5.3-Flash share?
35 benchmarks have published results for both models. GLM-5.2 has 51 scored results on Noometry and GLM-5.3-Flash has 40.