Model comparison
GLM-4.5-Air vs GPT-4.1
GLM-4.5-Air is the stronger model overall, scoring 38.7 to 35.7 on the Noometry Index.
Last verified . 26 shared benchmarks.
Summary
- They share 26 benchmarks with published results for both. GLM-4.5-Air scores higher in 4 categories and GPT-4.1 in 4 categories; 6 gaps are clear of the uncertainty.
- The widest gap is in math, where GLM-4.5-Air leads 33.8 to 21.4.
- The biggest single-benchmark swing is Kagi LLM Benchmark: 43% for GLM-4.5-Air and 52.3% for GPT-4.1.
- GLM-4.5-Air is cheaper at $0.20 / $1.10 per million input/output tokens, against $2 / $8 for GPT-4.1.
- GPT-4.1 accepts more context: 1.05M tokens versus 131K.
- GLM-4.5-Air has downloadable open weights; the other is API-only.
Side by side
| GLM-4.5-Air | GPT-4.1 | |
|---|---|---|
| Provider | Z.ai (Zhipu) | OpenAI |
| Noometry Index | 38.7 | 35.7 |
| Released | 2025-07-20 | 2025-04-14 |
| Weights | Open | Proprietary |
| Context window | 131K | 1.05M |
| Max output | 98K | 33K |
| Input $ / M tokens | $0.20 | $2 |
| Output $ / M tokens | $1.10 | $8 |
| Results tracked | 27 | 52 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Too close to call
GLM-4.5-Air: 34.5 (#225), GPT-4.1: 34.2 (#228)
| Benchmark | GLM-4.5-Air | GPT-4.1 |
|---|---|---|
| LMArena Coding | 1397 | 1391 |
| SWE-bench Verified | — | 48.5% |
| SWE-bench Verified (bash only) | — | 39.6% |
| Aider Polyglot | — | 52.4% |
| GSO | 2.9% | — |
| WeirdML | — | 39% |
| CadEval | — | 42% |
| ALE-Bench | — | 558.1 |
Agentic & Tool Use Not comparable
GLM-4.5-Air: —, GPT-4.1: 34.4 (#40)
| Benchmark | GLM-4.5-Air | GPT-4.1 |
|---|---|---|
| Berkeley Function Calling Leaderboard | — | 54% |
Reasoning GLM-4.5-Air leads
GLM-4.5-Air: 23.3 (#166), GPT-4.1: 11.7 (#334)
| Benchmark | GLM-4.5-Air | GPT-4.1 |
|---|---|---|
| Kagi LLM Benchmark | 43% | 52.3% |
| LMArena Hard Prompts | 1379 | 1384 |
| ForecastBench | 59.2 | 61.5 |
| ARC-AGI-2 | — | 0.4% |
| SimpleBench | — | 27% |
| ARC-AGI-1 | — | 5.5% |
| Chess Puzzles | — | 6% |
| EnigmaEval | — | 2.2% |
| DTBench | — | 68.3% |
| LMCA | — | 25.6% |
| Epoch Capabilities Index | — | 136.79 |
Math GLM-4.5-Air leads
GLM-4.5-Air: 33.8 (#166), GPT-4.1: 21.4 (#269)
| Benchmark | GLM-4.5-Air | GPT-4.1 |
|---|---|---|
| Omni-MATH | 39.1% | 47.1% |
| LMArena Math | 1396 | 1369 |
| FrontierMath (Tiers 1-3) | — | 6% |
| OTIS Mock AIME 2024-2025 | — | 38.3% |
| MATH Level 5 | — | 83% |
| FrontierMath (Feb 2025 set) | — | 5.5% |
| FrontierMath Tier 4 (v1) | — | 0% |
Knowledge GPT-4.1 leads
GLM-4.5-Air: 34.6 (#184), GPT-4.1: 37.0 (#152)
| Benchmark | GLM-4.5-Air | GPT-4.1 |
|---|---|---|
| Humanity's Last Exam | 8.1% | 5.4% |
| MMLU-Pro | 76.2% | 81.1% |
| Vectara Hallucination Rate | 9.3% | 5.6% |
| GPQA (HELM) | 59.4% | 65.9% |
| LMArena Expert | 1370 | 1364 |
| GPQA Diamond | — | 66.9% |
| SimpleQA Verified | — | 31.1% |
Multimodal Not comparable
GLM-4.5-Air: —, GPT-4.1: 37.5 (#66)
| Benchmark | GLM-4.5-Air | GPT-4.1 |
|---|---|---|
| LMArena Vision | — | 1210 |
| GeoBench | — | 72% |
Multilingual Too close to call
GLM-4.5-Air: 49.4 (#130), GPT-4.1: 49.6 (#128)
| Benchmark | GLM-4.5-Air | GPT-4.1 |
|---|---|---|
| LMArena Non-English | 1366 | 1370 |
| LMArena Chinese | 1427 | 1382 |
| LMArena French | 1399 | 1383 |
| LMArena German | 1376 | 1380 |
| LMArena Japanese | 1348 | 1320 |
| LMArena Korean | 1306 | 1339 |
| LMArena Russian | 1373 | 1377 |
| LMArena Spanish | 1385 | 1375 |
Instruction Following GPT-4.1 leads
GLM-4.5-Air: 69.8 (#165), GPT-4.1: 71.5 (#148)
| Benchmark | GLM-4.5-Air | GPT-4.1 |
|---|---|---|
| IFEval | 81.2% | 83.8% |
| LMArena Instruction Following | 1353 | 1367 |
Long Context GLM-4.5-Air leads
GLM-4.5-Air: 41.6 (#130), GPT-4.1: 40.0 (#158)
| Benchmark | GLM-4.5-Air | GPT-4.1 |
|---|---|---|
| LMArena Longer Query | 1367 | 1385 |
| Fiction.LiveBench | — | 63.9% |
Writing & Preference GPT-4.1 leads
GLM-4.5-Air: 56.1 (#134), GPT-4.1: 57.8 (#120)
| Benchmark | GLM-4.5-Air | GPT-4.1 |
|---|---|---|
| LMArena Text | 1384 | 1383 |
| LMArena Creative Writing | 1343 | 1363 |
| WildBench | 78.9% | 85.4% |
| LMArena Multi-Turn | 1371 | 1398 |
| EQ-Bench Creative Writing | — | 1420 |
Frequently asked questions
Is GLM-4.5-Air better than GPT-4.1?
GLM-4.5-Air is the stronger model overall, scoring 38.7 to 35.7 on the Noometry Index.
Which is cheaper, GLM-4.5-Air or GPT-4.1?
GLM-4.5-Air is cheaper. It lists at $0.20 per million input tokens and $1.10 per million output tokens; GPT-4.1 lists at $2 and $8.
Is GLM-4.5-Air or GPT-4.1 better for coding?
They score almost the same on coding (34.5 vs 34.2); test both on your own repository before choosing.
Which has the bigger context window?
GPT-4.1 does, with 1.05M tokens against 131K.
How many benchmarks do GLM-4.5-Air and GPT-4.1 share?
26 benchmarks have published results for both models. GLM-4.5-Air has 27 scored results on Noometry and GPT-4.1 has 52.