Model comparison
GLM-5.3 vs Qwen2.5-Max
GLM-5.3 is the stronger model overall, scoring 54.8 to 40.7 on the Noometry Index.
Last verified . 18 shared benchmarks.
Alibaba (Qwen)Summary
- They share 18 benchmarks with published results for both. GLM-5.3 scores higher in 8 categories and Qwen2.5-Max in 0 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in math, where GLM-5.3 leads 62.3 to 36.9.
- GLM-5.3 has downloadable open weights; the other is API-only.
Side by side
| GLM-5.3 | Qwen2.5-Max | |
|---|---|---|
| Provider | Z.ai (Zhipu) | Alibaba (Qwen) |
| Noometry Index | 54.8 | 40.7 |
| Released | 2026-08-14 | 2025-01-25 |
| Weights | Open | Proprietary |
| Context window | 1M | — |
| Max output | 131K | — |
| Input $ / M tokens | $1.40 | — |
| Output $ / M tokens | $4.40 | — |
| Results tracked | 42 | 27 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding GLM-5.3 leads
GLM-5.3: 59.5 (#14), Qwen2.5-Max: 41.8 (#117)
| Benchmark | GLM-5.3 | Qwen2.5-Max |
|---|---|---|
| LMArena Coding | 1496 | 1359 |
| DeepSWE | 69% | — |
| FrontierCode | 40.1% | — |
| CursorBench | 42.6% | — |
| LMArena WebDev | 1622 | — |
| FrontierSWE | 30.2% | — |
| SciCode | 59% | — |
| WeirdML | 75.4% | — |
| LiveBench Coding | — | 64.4% |
| ALE-Bench | 1,317 | — |
Agentic & Tool Use Not comparable
GLM-5.3: 36.4 (#38), Qwen2.5-Max: —
| Benchmark | GLM-5.3 | Qwen2.5-Max |
|---|---|---|
| APEX-Agents | 56.6% | — |
| Vending-Bench 2 | 8,164 | — |
Reasoning GLM-5.3 leads
GLM-5.3: 46.1 (#46), Qwen2.5-Max: 25.6 (#147)
| Benchmark | GLM-5.3 | Qwen2.5-Max |
|---|---|---|
| LMArena Hard Prompts | 1489 | 1360 |
| Epoch Capabilities Index | 155.61 | 132.53 |
| NYT Connections (extended) | 74.2% | — |
| CritPt | 19.1% | — |
| Chess Puzzles | 21% | — |
| LiveBench Reasoning | — | 51.4% |
| Mystery Game Puzzles | 33% | — |
| DTBench | 87.7% | — |
| LiveBench Data Analysis | — | 67.9% |
| LMCA | 55.5% | — |
| Bench to the Future 3 | 0.15 | — |
| LiveBench | — | 62.3% |
Math GLM-5.3 leads
GLM-5.3: 62.3 (#33), Qwen2.5-Max: 36.9 (#162)
| Benchmark | GLM-5.3 | Qwen2.5-Max |
|---|---|---|
| LMArena Math | 1489 | 1369 |
| FrontierMath (Tiers 1-3) | 68.8% | — |
| FrontierMath Tier 4 | 29.3% | — |
| OTIS Mock AIME 2024-2025 | 91.1% | — |
| ProofBench | 49% | — |
| LiveBench Math | — | 58.4% |
Knowledge GLM-5.3 leads
GLM-5.3: 58.3 (#37), Qwen2.5-Max: 35.3 (#186)
| Benchmark | GLM-5.3 | Qwen2.5-Max |
|---|---|---|
| LMArena Expert | 1516 | 1337 |
| GPQA Diamond | 90.9% | — |
| SimpleQA Verified | 41% | — |
| Confabulations | — | 21.8% |
Multilingual GLM-5.3 leads
GLM-5.3: 55.7 (#28), Qwen2.5-Max: 48.1 (#146)
| Benchmark | GLM-5.3 | Qwen2.5-Max |
|---|---|---|
| LMArena Non-English | 1457 | 1352 |
| LMArena Chinese | 1528 | 1382 |
| LMArena French | 1499 | 1396 |
| LMArena German | 1499 | 1350 |
| LMArena Japanese | 1453 | 1300 |
| LMArena Korean | 1472 | 1304 |
| LMArena Russian | 1463 | 1353 |
| LMArena Spanish | 1460 | 1377 |
Instruction Following GLM-5.3 leads
GLM-5.3: 77.5 (#23), Qwen2.5-Max: 71.3 (#152)
| Benchmark | GLM-5.3 | Qwen2.5-Max |
|---|---|---|
| LMArena Instruction Following | 1477 | 1335 |
| LiveBench Instruction Following | — | 75.3% |
Long Context GLM-5.3 leads
GLM-5.3: 45.4 (#41), Qwen2.5-Max: 41.4 (#142)
| Benchmark | GLM-5.3 | Qwen2.5-Max |
|---|---|---|
| LMArena Longer Query | 1482 | 1358 |
Writing & Preference GLM-5.3 leads
GLM-5.3: 75.7 (#6), Qwen2.5-Max: 55.4 (#146)
| Benchmark | GLM-5.3 | Qwen2.5-Max |
|---|---|---|
| LMArena Text | 1471 | 1367 |
| LMArena Creative Writing | 1457 | 1339 |
| LMArena Multi-Turn | 1472 | 1364 |
| Short-Story Creative Writing | — | 72.9% |
| EQ-Bench Creative Writing | 2075 | — |
| LiveBench Language | — | 56.3% |
Frequently asked questions
Is GLM-5.3 better than Qwen2.5-Max?
GLM-5.3 is the stronger model overall, scoring 54.8 to 40.7 on the Noometry Index.
Is GLM-5.3 or Qwen2.5-Max better for coding?
GLM-5.3 scores higher on coding benchmarks: 59.5 versus 41.8 in the Noometry coding category.
How many benchmarks do GLM-5.3 and Qwen2.5-Max share?
18 benchmarks have published results for both models. GLM-5.3 has 42 scored results on Noometry and Qwen2.5-Max has 27.