Model comparison
GLM-5 vs GPT-5.4 mini
GLM-5 is the stronger model overall, scoring 45.9 to 44.8 on the Noometry Index.
Last verified . 33 shared benchmarks.
Summary
- They share 33 benchmarks with published results for both. GLM-5 scores higher in 7 categories and GPT-5.4 mini in 2 categories; 7 gaps are clear of the uncertainty.
- The widest gap is in coding, where GLM-5 leads 49.0 to 45.0.
- The biggest single-benchmark swing is Kagi LLM Benchmark: 75% for GLM-5 and 37.9% for GPT-5.4 mini.
- GLM-5 is cheaper at $1 / $3.20 per million input/output tokens, against $0.75 / $4.50 for GPT-5.4 mini.
- GPT-5.4 mini accepts more context: 400K tokens versus 205K.
- GLM-5 has downloadable open weights; the other is API-only.
Side by side
| GLM-5 | GPT-5.4 mini | |
|---|---|---|
| Provider | Z.ai (Zhipu) | OpenAI |
| Noometry Index | 45.9 | 44.8 |
| Released | 2026-02-11 | 2026-03-17 |
| Weights | Open | Proprietary |
| Context window | 205K | 400K |
| Max output | 131K | 128K |
| Input $ / M tokens | $1 | $0.75 |
| Output $ / M tokens | $3.20 | $4.50 |
| Results tracked | 45 | 46 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding GLM-5 leads
GLM-5: 49.0 (#47), GPT-5.4 mini: 45.0 (#67)
| Benchmark | GLM-5 | GPT-5.4 mini |
|---|---|---|
| LMArena WebDev | 1434 | 1397 |
| WeirdML | 48.2% | 60.3% |
| LMArena Coding | 1461 | 1438 |
| ALE-Bench | 765.62 | 1,189 |
| SWE-bench Verified | 72.1% | — |
| FrontierCode | — | 27% |
| SWE-bench Verified (bash only) | 72.8% | — |
| SWE-bench Multilingual | 69.7% | — |
| SciCode | — | 49.9% |
Agentic & Tool Use GLM-5 leads
GLM-5: 30.9 (#65), GPT-5.4 mini: 29.6 (#74)
| Benchmark | GLM-5 | GPT-5.4 mini |
|---|---|---|
| Terminal-Bench | 52.4% | — |
| τ²-bench Airline | 82.5% | — |
| τ²-bench Banking | 9.8% | — |
| τ²-bench Retail | 73.7% | — |
| τ²-bench Telecom | 86.8% | — |
| DeepResearch Bench | — | 36.3% |
| Vending-Bench 2 | 4,432 | — |
Reasoning GPT-5.4 mini leads
GLM-5: 27.6 (#100), GPT-5.4 mini: 30.3 (#77)
| Benchmark | GLM-5 | GPT-5.4 mini |
|---|---|---|
| ARC-AGI-2 | 4.9% | 18.9% |
| Kagi LLM Benchmark | 75% | 37.9% |
| NYT Connections (extended) | 74.8% | 61.8% |
| ARC-AGI-1 | 44.7% | 63.7% |
| Chess Puzzles | 10% | 24% |
| LMArena Hard Prompts | 1451 | 1424 |
| Epoch Capabilities Index | 145.84 | 148.83 |
| ForecastBench | 61 | 57 |
| SimpleBench | 53.2% | — |
| CritPt | — | 10% |
| Thematic Generalization | — | 61.7% |
| Mystery Game Puzzles | — | 11% |
| DTBench | — | 80% |
| LMCA | — | 40.8% |
Math Too close to call
GLM-5: 44.8 (#66), GPT-5.4 mini: 45.4 (#63)
| Benchmark | GLM-5 | GPT-5.4 mini |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 80% | 88.9% |
| LMArena Math | 1440 | 1418 |
| FrontierMath (Feb 2025 set) | 16.4% | 28.3% |
| FrontierMath Tier 4 (v1) | 2.1% | 2.1% |
| FrontierMath (Tiers 1-3) | — | 51.2% |
| FrontierMath Tier 4 | — | 9.8% |
| MathArena Final-Answer Competitions | 65.7% | — |
| ProofBench | — | 21% |
Knowledge Too close to call
GLM-5: 52.2 (#59), GPT-5.4 mini: 51.4 (#62)
| Benchmark | GLM-5 | GPT-5.4 mini |
|---|---|---|
| GPQA Diamond | 87.8% | 86.9% |
| Vectara Hallucination Rate | 10.1% | 5.5% |
| LMArena Expert | 1454 | 1434 |
| SimpleQA Verified | — | 29.4% |
Multimodal Not comparable
GLM-5: —, GPT-5.4 mini: 39.0 (#51)
| Benchmark | GLM-5 | GPT-5.4 mini |
|---|---|---|
| LMArena Vision | — | 1245 |
Multilingual GLM-5 leads
GLM-5: 53.9 (#57), GPT-5.4 mini: 52.1 (#94)
| Benchmark | GLM-5 | GPT-5.4 mini |
|---|---|---|
| LMArena Non-English | 1430 | 1405 |
| LMArena Chinese | 1511 | 1446 |
| LMArena French | 1455 | 1440 |
| LMArena German | 1445 | 1407 |
| LMArena Japanese | 1416 | 1375 |
| LMArena Korean | 1421 | 1369 |
| LMArena Russian | 1436 | 1417 |
| LMArena Spanish | 1452 | 1403 |
Instruction Following GLM-5 leads
GLM-5: 75.4 (#64), GPT-5.4 mini: 74.3 (#99)
| Benchmark | GLM-5 | GPT-5.4 mini |
|---|---|---|
| LMArena Instruction Following | 1427 | 1405 |
Long Context GLM-5 leads
GLM-5: 44.6 (#58), GPT-5.4 mini: 42.9 (#107)
| Benchmark | GLM-5 | GPT-5.4 mini |
|---|---|---|
| LMArena Longer Query | 1446 | 1407 |
| CL-bench | 18.7% | — |
Writing & Preference GLM-5 leads
GLM-5: 66.2 (#38), GPT-5.4 mini: 64.2 (#56)
| Benchmark | GLM-5 | GPT-5.4 mini |
|---|---|---|
| LMArena Text | 1446 | 1411 |
| LMArena Creative Writing | 1438 | 1370 |
| EQ-Bench Creative Writing | 1601 | 1665 |
| LMArena Multi-Turn | 1456 | 1429 |
Frequently asked questions
Is GLM-5 better than GPT-5.4 mini?
GLM-5 is the stronger model overall, scoring 45.9 to 44.8 on the Noometry Index.
Which is cheaper, GLM-5 or GPT-5.4 mini?
GLM-5 is cheaper. It lists at $1 per million input tokens and $3.20 per million output tokens; GPT-5.4 mini lists at $0.75 and $4.50.
Is GLM-5 or GPT-5.4 mini better for coding?
GLM-5 scores higher on coding benchmarks: 49.0 versus 45.0 in the Noometry coding category.
Which has the bigger context window?
GPT-5.4 mini does, with 400K tokens against 205K.
How many benchmarks do GLM-5 and GPT-5.4 mini share?
33 benchmarks have published results for both models. GLM-5 has 45 scored results on Noometry and GPT-5.4 mini has 46.