Model comparison
GLM-5.3 vs Qwen3.7 Max
GLM-5.3 is the stronger model overall, scoring 54.8 to 51.5 on the Noometry Index.
Last verified . 28 shared benchmarks.
Summary
- They share 28 benchmarks with published results for both. GLM-5.3 scores higher in 4 categories and Qwen3.7 Max in 5 categories; 6 gaps are clear of the uncertainty.
- The widest gap is in agentic & tool use, where GLM-5.3 leads 36.4 to 22.1.
- The biggest single-benchmark swing is ProofBench: 49% for GLM-5.3 and 26% for Qwen3.7 Max.
- GLM-5.3 is cheaper at $1.40 / $4.40 per million input/output tokens, against $2.50 / $7.50 for Qwen3.7 Max.
- GLM-5.3 has downloadable open weights; the other is API-only.
Side by side
| GLM-5.3 | Qwen3.7 Max | |
|---|---|---|
| Provider | Z.ai (Zhipu) | Alibaba (Qwen) |
| Noometry Index | 54.8 | 51.5 |
| Released | 2026-08-14 | 2026-05-19 |
| Weights | Open | Proprietary |
| Context window | 1M | 1M |
| Max output | 131K | 131K |
| Input $ / M tokens | $1.40 | $2.50 |
| Output $ / M tokens | $4.40 | $7.50 |
| Results tracked | 42 | 33 |
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), Qwen3.7 Max: 50.4 (#45)
| Benchmark | GLM-5.3 | Qwen3.7 Max |
|---|---|---|
| LMArena WebDev | 1622 | 1515 |
| SciCode | 59% | 48.8% |
| LMArena Coding | 1496 | 1498 |
| ALE-Bench | 1,317 | 1,189 |
| SWE-bench Verified | — | 77.3% |
| DeepSWE | 69% | — |
| FrontierCode | 40.1% | — |
| CursorBench | 42.6% | — |
| FrontierSWE | 30.2% | — |
| WeirdML | 75.4% | — |
Agentic & Tool Use GLM-5.3 leads
GLM-5.3: 36.4 (#38), Qwen3.7 Max: 22.1 (#135)
| Benchmark | GLM-5.3 | Qwen3.7 Max |
|---|---|---|
| APEX-Agents | 56.6% | — |
| GBAEval | — | 0.4% |
| Vending-Bench 2 | 8,164 | — |
Reasoning Qwen3.7 Max leads
GLM-5.3: 46.1 (#46), Qwen3.7 Max: 49.2 (#38)
| Benchmark | GLM-5.3 | Qwen3.7 Max |
|---|---|---|
| NYT Connections (extended) | 74.2% | 85.1% |
| CritPt | 19.1% | 13.4% |
| Chess Puzzles | 21% | 19% |
| LMArena Hard Prompts | 1489 | 1483 |
| Mystery Game Puzzles | 33% | 32% |
| DTBench | 87.7% | 92.3% |
| LMCA | 55.5% | 44% |
| Epoch Capabilities Index | 155.61 | 153.68 |
| SimpleBench | — | 70.4% |
| EBR-Bench | — | 9.5% |
| Bench to the Future 3 | 0.15 | — |
Math Too close to call
GLM-5.3: 62.3 (#33), Qwen3.7 Max: 62.4 (#32)
| Benchmark | GLM-5.3 | Qwen3.7 Max |
|---|---|---|
| FrontierMath (Tiers 1-3) | 68.8% | 64.6% |
| FrontierMath Tier 4 | 29.3% | 34.1% |
| OTIS Mock AIME 2024-2025 | 91.1% | 95.6% |
| ProofBench | 49% | 26% |
| LMArena Math | 1489 | 1490 |
Knowledge Qwen3.7 Max leads
GLM-5.3: 58.3 (#37), Qwen3.7 Max: 61.6 (#28)
| Benchmark | GLM-5.3 | Qwen3.7 Max |
|---|---|---|
| GPQA Diamond | 90.9% | 90.9% |
| SimpleQA Verified | 41% | 55.8% |
| LMArena Expert | 1516 | 1488 |
Multilingual Qwen3.7 Max leads
GLM-5.3: 55.7 (#28), Qwen3.7 Max: 56.9 (#15)
| Benchmark | GLM-5.3 | Qwen3.7 Max |
|---|---|---|
| LMArena Non-English | 1457 | 1474 |
| LMArena Chinese | 1528 | 1530 |
| LMArena Russian | 1463 | 1484 |
| LMArena French | 1499 | — |
| LMArena German | 1499 | — |
| LMArena Japanese | 1453 | — |
| LMArena Korean | 1472 | — |
| LMArena Spanish | 1460 | — |
Instruction Following Too close to call
GLM-5.3: 77.5 (#23), Qwen3.7 Max: 76.7 (#38)
| Benchmark | GLM-5.3 | Qwen3.7 Max |
|---|---|---|
| LMArena Instruction Following | 1477 | 1460 |
Long Context Too close to call
GLM-5.3: 45.4 (#41), Qwen3.7 Max: 45.4 (#40)
| Benchmark | GLM-5.3 | Qwen3.7 Max |
|---|---|---|
| LMArena Longer Query | 1482 | 1482 |
Writing & Preference GLM-5.3 leads
GLM-5.3: 75.7 (#6), Qwen3.7 Max: 65.0 (#54)
| Benchmark | GLM-5.3 | Qwen3.7 Max |
|---|---|---|
| LMArena Text | 1471 | 1476 |
| LMArena Creative Writing | 1457 | 1449 |
| LMArena Multi-Turn | 1472 | 1481 |
| EQ-Bench Creative Writing | 2075 | — |
| EQ-Bench 4 | — | 1110 |
Frequently asked questions
Is GLM-5.3 better than Qwen3.7 Max?
GLM-5.3 is the stronger model overall, scoring 54.8 to 51.5 on the Noometry Index.
Which is cheaper, GLM-5.3 or Qwen3.7 Max?
GLM-5.3 is cheaper. It lists at $1.40 per million input tokens and $4.40 per million output tokens; Qwen3.7 Max lists at $2.50 and $7.50.
Is GLM-5.3 or Qwen3.7 Max better for coding?
GLM-5.3 scores higher on coding benchmarks: 59.5 versus 50.4 in the Noometry coding category.
Which has the bigger context window?
Both accept 1M tokens.
How many benchmarks do GLM-5.3 and Qwen3.7 Max share?
28 benchmarks have published results for both models. GLM-5.3 has 42 scored results on Noometry and Qwen3.7 Max has 33.