Model comparison
GLM-5 vs Qwen2.5-Coder-32B
GLM-5 is the stronger model overall, scoring 46.1 to 33.4 on the Noometry Index. Qwen2.5-Coder-32B costs 2.1× less per token, which makes it the better buy when GLM-5's lead doesn't matter for your workload.
Last verified . 14 shared benchmarks.
Alibaba (Qwen)Summary
- They share 14 benchmarks with published results for both. GLM-5 scores higher in 8 categories and Qwen2.5-Coder-32B in 0 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in coding, where GLM-5 leads 49.0 to 22.6.
- The biggest single-benchmark swing is SWE-bench Verified (bash only): 72.8% for GLM-5 and 9% for Qwen2.5-Coder-32B.
- Qwen2.5-Coder-32B is cheaper at $0.66 / $1 per million input/output tokens, against $1 / $3.20 for GLM-5.
- GLM-5 accepts more context: 205K tokens versus 33K.
Side by side
| GLM-5 | Qwen2.5-Coder-32B | |
|---|---|---|
| Provider | Z.ai (Zhipu) | Alibaba (Qwen) |
| Noometry Index | 46.1 | 33.4 |
| Released | 2026-02-11 | 2024-09-18 |
| Weights | Open | Open |
| Context window | 205K | 33K |
| Max output | 131K | 29K |
| Input $ / M tokens | $1 | $0.66 |
| Output $ / M tokens | $3.20 | $1 |
| Results tracked | 45 | 31 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding GLM-5 leads
GLM-5: 49.0 (#52), Qwen2.5-Coder-32B: 22.6 (#333)
| Benchmark | GLM-5 | Qwen2.5-Coder-32B |
|---|---|---|
| SWE-bench Verified (bash only) | 72.8% | 9% |
| LMArena Coding | 1461 | 1276 |
| SWE-bench Verified | 72.1% | — |
| Aider Polyglot | — | 16.4% |
| LMArena WebDev | 1434 | — |
| SWE-bench Multilingual | 69.7% | — |
| WeirdML | 48.2% | — |
| BigCodeBench Instruct | — | 49% |
| LiveBench Coding | — | 56.9% |
| BigCodeBench Complete | — | 58% |
| ALE-Bench | 765.62 | — |
| HumanEval+ | — | 87.2% |
| MBPP+ | — | 77% |
Agentic & Tool Use Not comparable
GLM-5: 31.1 (#71), Qwen2.5-Coder-32B: —
| Benchmark | GLM-5 | Qwen2.5-Coder-32B |
|---|---|---|
| Terminal-Bench | 52.4% | — |
| τ²-bench Airline | 82.5% | — |
| τ²-bench Banking | 9.8% | — |
| τ²-bench Retail | 73.7% | — |
| τ²-bench Telecom | 86.8% | — |
| Vending-Bench 2 | 4,432 | — |
Reasoning GLM-5 leads
GLM-5: 27.6 (#116), Qwen2.5-Coder-32B: 21.2 (#225)
| Benchmark | GLM-5 | Qwen2.5-Coder-32B |
|---|---|---|
| LMArena Hard Prompts | 1452 | 1251 |
| Epoch Capabilities Index | 145.83 | 119.49 |
| ARC-AGI-2 | 4.9% | — |
| SimpleBench | 53.2% | — |
| Kagi LLM Benchmark | 75% | — |
| NYT Connections (extended) | 74.8% | — |
| ARC-AGI-1 | 44.7% | — |
| Chess Puzzles | 10% | — |
| LiveBench Reasoning | — | 42.1% |
| LiveBench Data Analysis | — | 49.9% |
| ForecastBench | 61 | — |
| HellaSwag | — | 83% |
| LiveBench | — | 46.2% |
| WinoGrande | — | 80.8% |
Math GLM-5 leads
GLM-5: 46.4 (#71), Qwen2.5-Coder-32B: 33.3 (#204)
| Benchmark | GLM-5 | Qwen2.5-Coder-32B |
|---|---|---|
| LMArena Math | 1440 | 1251 |
| MathArena Final-Answer Competitions | 65.7% | — |
| OTIS Mock AIME 2024-2025 | 80% | — |
| LiveBench Math | — | 46.6% |
| FrontierMath (Feb 2025 set) | 16.4% | — |
| FrontierMath Tier 4 (v1) | 2.1% | — |
| GSM8K | — | 93% |
Knowledge GLM-5 leads
GLM-5: 52.3 (#64), Qwen2.5-Coder-32B: 33.4 (#203)
| Benchmark | GLM-5 | Qwen2.5-Coder-32B |
|---|---|---|
| LMArena Expert | 1454 | 1221 |
| GPQA Diamond | 87.8% | — |
| Vectara Hallucination Rate | 10.1% | — |
| ARC (AI2) Challenge | — | 70.5% |
| MMLU | — | 79.1% |
Multilingual GLM-5 leads
GLM-5: 53.7 (#58), Qwen2.5-Coder-32B: 37.8 (#235)
| Benchmark | GLM-5 | Qwen2.5-Coder-32B |
|---|---|---|
| LMArena Non-English | 1430 | 1205 |
| LMArena Chinese | 1511 | 1222 |
| LMArena Russian | 1436 | 1228 |
| LMArena French | 1455 | — |
| LMArena German | 1445 | — |
| LMArena Japanese | 1416 | — |
| LMArena Korean | 1423 | — |
| LMArena Spanish | 1454 | — |
Instruction Following GLM-5 leads
GLM-5: 75.2 (#67), Qwen2.5-Coder-32B: 61.4 (#245)
| Benchmark | GLM-5 | Qwen2.5-Coder-32B |
|---|---|---|
| LMArena Instruction Following | 1428 | 1223 |
| LiveBench Instruction Following | — | 58.7% |
Long Context GLM-5 leads
GLM-5: 44.7 (#60), Qwen2.5-Coder-32B: 38.0 (#208)
| Benchmark | GLM-5 | Qwen2.5-Coder-32B |
|---|---|---|
| LMArena Longer Query | 1446 | 1251 |
| CL-bench | 18.7% | — |
Writing & Preference GLM-5 leads
GLM-5: 66.0 (#38), Qwen2.5-Coder-32B: 41.6 (#240)
| Benchmark | GLM-5 | Qwen2.5-Coder-32B |
|---|---|---|
| LMArena Text | 1446 | 1230 |
| LMArena Creative Writing | 1439 | 1174 |
| LMArena Multi-Turn | 1456 | 1222 |
| EQ-Bench Creative Writing | 1601 | — |
| LiveBench Language | — | 23.3% |
Frequently asked questions
Is GLM-5 better than Qwen2.5-Coder-32B?
GLM-5 is the stronger model overall, scoring 46.1 to 33.4 on the Noometry Index. Qwen2.5-Coder-32B costs 2.1× less per token, which makes it the better buy when GLM-5's lead doesn't matter for your workload.
Which is cheaper, GLM-5 or Qwen2.5-Coder-32B?
Qwen2.5-Coder-32B is cheaper. It lists at $0.66 per million input tokens and $1 per million output tokens; GLM-5 lists at $1 and $3.20.
Is GLM-5 or Qwen2.5-Coder-32B better for coding?
GLM-5 scores higher on coding benchmarks: 49.0 versus 22.6 in the Noometry coding category.
Which has the bigger context window?
GLM-5 does, with 205K tokens against 33K.
How many benchmarks do GLM-5 and Qwen2.5-Coder-32B share?
14 benchmarks have published results for both models. GLM-5 has 45 scored results on Noometry and Qwen2.5-Coder-32B has 31.