Model comparison
GLM-5.3-Flash vs Qwen3.8 Max
Qwen3.8 Max is the stronger model overall, scoring 56.8 to 51.8 on the Noometry Index. GLM-5.3-Flash costs 13× less per token, which makes it the better buy when Qwen3.8 Max's lead doesn't matter for your workload.
Last verified . 33 shared benchmarks.
Alibaba (Qwen)Summary
- They share 33 benchmarks with published results for both. GLM-5.3-Flash scores higher in 1 category and Qwen3.8 Max in 9 categories; 6 gaps are clear of the uncertainty.
- The widest gap is in math, where Qwen3.8 Max leads 73.2 to 53.3.
- The biggest single-benchmark swing is ProofBench: 21% for GLM-5.3-Flash and 58% for Qwen3.8 Max.
- GLM-5.3-Flash is cheaper at $0.15 / $0.50 per million input/output tokens, against $2 / $6 for Qwen3.8 Max.
- GLM-5.3-Flash has downloadable open weights; the other is API-only.
Side by side
| GLM-5.3-Flash | Qwen3.8 Max | |
|---|---|---|
| Provider | Z.ai (Zhipu) | Alibaba (Qwen) |
| Noometry Index | 51.8 | 56.8 |
| Released | 2026-08-20 | 2026-08-02 |
| Weights | Open | Proprietary |
| Context window | 1M | 1M |
| Max output | 131K | 131K |
| Input $ / M tokens | $0.15 | $2 |
| Output $ / M tokens | $0.50 | $6 |
| Results tracked | 40 | 39 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Too close to call
GLM-5.3-Flash: 53.1 (#31), Qwen3.8 Max: 53.5 (#29)
| Benchmark | GLM-5.3-Flash | Qwen3.8 Max |
|---|---|---|
| DeepSWE | 63.4% | 57.5% |
| LMArena WebDev | 1609 | 1674 |
| FrontierSWE | 18.1% | 17.8% |
| SciCode | 51.6% | 53.2% |
| LMArena Coding | 1508 | 1502 |
| FrontierCode | 31.8% | — |
| CursorBench | 36.8% | — |
| ALE-Bench | 303.55 | — |
Agentic & Tool Use Qwen3.8 Max leads
GLM-5.3-Flash: 34.2 (#47), Qwen3.8 Max: 45.4 (#14)
| Benchmark | GLM-5.3-Flash | Qwen3.8 Max |
|---|---|---|
| APEX-Agents | 52.8% | 63.3% |
| GDP.pdf | 14% | 23.2% |
| τ²-bench Banking | — | 55.1% |
Reasoning Qwen3.8 Max leads
GLM-5.3-Flash: 48.0 (#42), Qwen3.8 Max: 54.4 (#26)
| Benchmark | GLM-5.3-Flash | Qwen3.8 Max |
|---|---|---|
| CritPt | 15.4% | 20% |
| Chess Puzzles | 14% | 40% |
| LMArena Hard Prompts | 1491 | 1496 |
| Mystery Game Puzzles | 8% | 38% |
| Epoch Capabilities Index | 151.88 | 156.41 |
| ARC-AGI-2 | 65.8% | — |
| NYT Connections (extended) | — | 88.3% |
| ARC-AGI-1 | 91% | — |
| DTBench | — | 92% |
| LMCA | — | 46.2% |
| Surface Evolver Bench | 52.5% | — |
| Bench to the Future 3 | 0.15 | — |
Math Qwen3.8 Max leads
GLM-5.3-Flash: 53.3 (#47), Qwen3.8 Max: 73.2 (#20)
| Benchmark | GLM-5.3-Flash | Qwen3.8 Max |
|---|---|---|
| FrontierMath (Tiers 1-3) | 55.8% | 74.7% |
| FrontierMath Tier 4 | 17.1% | 46.3% |
| OTIS Mock AIME 2024-2025 | 93.9% | 100% |
| ProofBench | 21% | 58% |
| LMArena Math | 1500 | 1499 |
Knowledge Qwen3.8 Max leads
GLM-5.3-Flash: 58.4 (#36), Qwen3.8 Max: 61.7 (#27)
| Benchmark | GLM-5.3-Flash | Qwen3.8 Max |
|---|---|---|
| GPQA Diamond | 90.2% | 92.7% |
| LMArena Expert | 1513 | 1507 |
| SimpleQA Verified | — | 47.3% |
Multimodal GLM-5.3-Flash leads
GLM-5.3-Flash: 42.8 (#27), Qwen3.8 Max: 37.2 (#75)
| Benchmark | GLM-5.3-Flash | Qwen3.8 Max |
|---|---|---|
| LMArena Vision | 1296 | 1314 |
| Furniture Assembly | — | 20% |
Multilingual Too close to call
GLM-5.3-Flash: 56.0 (#25), Qwen3.8 Max: 56.7 (#18)
| Benchmark | GLM-5.3-Flash | Qwen3.8 Max |
|---|---|---|
| LMArena Non-English | 1462 | 1472 |
| LMArena Chinese | 1527 | 1538 |
| LMArena French | 1496 | 1503 |
| LMArena German | 1470 | 1483 |
| LMArena Japanese | 1429 | 1467 |
| LMArena Korean | 1446 | 1461 |
| LMArena Russian | 1469 | 1481 |
| LMArena Spanish | 1471 | 1492 |
Instruction Following Too close to call
GLM-5.3-Flash: 77.5 (#20), Qwen3.8 Max: 77.6 (#17)
| Benchmark | GLM-5.3-Flash | Qwen3.8 Max |
|---|---|---|
| LMArena Instruction Following | 1478 | 1479 |
Long Context Too close to call
GLM-5.3-Flash: 45.4 (#39), Qwen3.8 Max: 45.6 (#31)
| Benchmark | GLM-5.3-Flash | Qwen3.8 Max |
|---|---|---|
| LMArena Longer Query | 1482 | 1489 |
Writing & Preference Qwen3.8 Max leads
GLM-5.3-Flash: 65.3 (#50), Qwen3.8 Max: 67.1 (#30)
| Benchmark | GLM-5.3-Flash | Qwen3.8 Max |
|---|---|---|
| LMArena Text | 1471 | 1483 |
| LMArena Creative Writing | 1442 | 1479 |
| LMArena Multi-Turn | 1467 | 1489 |
Frequently asked questions
Is GLM-5.3-Flash better than Qwen3.8 Max?
Qwen3.8 Max is the stronger model overall, scoring 56.8 to 51.8 on the Noometry Index. GLM-5.3-Flash costs 13× less per token, which makes it the better buy when Qwen3.8 Max's lead doesn't matter for your workload.
Which is cheaper, GLM-5.3-Flash or Qwen3.8 Max?
GLM-5.3-Flash is cheaper. It lists at $0.15 per million input tokens and $0.50 per million output tokens; Qwen3.8 Max lists at $2 and $6.
Is GLM-5.3-Flash or Qwen3.8 Max better for coding?
They score almost the same on coding (53.1 vs 53.5); test both on your own repository before choosing.
Which has the bigger context window?
Both accept 1M tokens.
How many benchmarks do GLM-5.3-Flash and Qwen3.8 Max share?
33 benchmarks have published results for both models. GLM-5.3-Flash has 40 scored results on Noometry and Qwen3.8 Max has 39.