Model comparison
GLM-4.6 vs GPT-5.4 mini
GPT-5.4 mini is the stronger model overall, scoring 45.0 to 41.4 on the Noometry Index. GLM-4.6 costs 1.7× less per token, which makes it the better buy when GPT-5.4 mini's lead doesn't matter for your workload.
Last verified . 26 shared benchmarks.
Summary
- They share 26 benchmarks with published results for both. GLM-4.6 scores higher in 4 categories and GPT-5.4 mini in 5 categories; 7 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where GPT-5.4 mini leads 51.5 to 40.2.
- The biggest single-benchmark swing is SciCode: 38.4% for GLM-4.6 and 49.9% for GPT-5.4 mini.
- GLM-4.6 is cheaper at $0.60 / $2.20 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-4.6 has downloadable open weights; the other is API-only.
Side by side
| GLM-4.6 | GPT-5.4 mini | |
|---|---|---|
| Provider | Z.ai (Zhipu) | OpenAI |
| Noometry Index | 41.4 | 45.0 |
| Released | 2025-09-30 | 2026-03-17 |
| Weights | Open | Proprietary |
| Context window | 205K | 400K |
| Max output | 131K | 128K |
| Input $ / M tokens | $0.60 | $0.75 |
| Output $ / M tokens | $2.20 | $4.50 |
| Results tracked | 29 | 46 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding GPT-5.4 mini leads
GLM-4.6: 40.1 (#148), GPT-5.4 mini: 45.2 (#72)
| Benchmark | GLM-4.6 | GPT-5.4 mini |
|---|---|---|
| LMArena WebDev | 1340 | 1397 |
| SciCode | 38.4% | 49.9% |
| LMArena Coding | 1449 | 1438 |
| ALE-Bench | 340.82 | 1,189 |
| FrontierCode | — | 27% |
| SWE-bench Verified (bash only) | 55.4% | — |
| WeirdML | — | 60.3% |
Agentic & Tool Use GLM-4.6 leads
GLM-4.6: 32.3 (#66), GPT-5.4 mini: 29.9 (#81)
| Benchmark | GLM-4.6 | GPT-5.4 mini |
|---|---|---|
| Terminal-Bench | 24.5% | — |
| Berkeley Function Calling Leaderboard | 72.4% | — |
| DeepResearch Bench | — | 36.3% |
Reasoning GPT-5.4 mini leads
GLM-4.6: 23.7 (#172), GPT-5.4 mini: 30.4 (#85)
| Benchmark | GLM-4.6 | GPT-5.4 mini |
|---|---|---|
| Kagi LLM Benchmark | 47.4% | 37.9% |
| CritPt | 1.1% | 10% |
| LMArena Hard Prompts | 1440 | 1424 |
| ARC-AGI-2 | — | 18.9% |
| NYT Connections (extended) | — | 61.8% |
| ARC-AGI-1 | — | 63.7% |
| Chess Puzzles | — | 24% |
| Thematic Generalization | — | 61.7% |
| Mystery Game Puzzles | — | 11% |
| DTBench | — | 80% |
| LMCA | — | 40.8% |
| Epoch Capabilities Index | — | 148.84 |
| ForecastBench | — | 57 |
Math GPT-5.4 mini leads
GLM-4.6: 39.1 (#111), GPT-5.4 mini: 45.5 (#75)
| Benchmark | GLM-4.6 | GPT-5.4 mini |
|---|---|---|
| LMArena Math | 1432 | 1419 |
| FrontierMath (Feb 2025 set) | 3.8% | 28.3% |
| FrontierMath Tier 4 (v1) | 2.1% | 2.1% |
| FrontierMath (Tiers 1-3) | — | 51.2% |
| FrontierMath Tier 4 | — | 9.8% |
| OTIS Mock AIME 2024-2025 | — | 88.9% |
| ProofBench | — | 21% |
Knowledge GPT-5.4 mini leads
GLM-4.6: 40.2 (#124), GPT-5.4 mini: 51.5 (#67)
| Benchmark | GLM-4.6 | GPT-5.4 mini |
|---|---|---|
| Vectara Hallucination Rate | 9.5% | 5.5% |
| LMArena Expert | 1431 | 1435 |
| GPQA Diamond | — | 86.9% |
| SimpleQA Verified | — | 29.4% |
Multimodal Not comparable
GLM-4.6: —, GPT-5.4 mini: 39.7 (#56)
| Benchmark | GLM-4.6 | GPT-5.4 mini |
|---|---|---|
| LMArena Vision | — | 1245 |
Multilingual GLM-4.6 leads
GLM-4.6: 53.5 (#66), GPT-5.4 mini: 51.9 (#96)
| Benchmark | GLM-4.6 | GPT-5.4 mini |
|---|---|---|
| LMArena Non-English | 1426 | 1405 |
| LMArena Chinese | 1499 | 1446 |
| LMArena French | 1459 | 1440 |
| LMArena German | 1447 | 1409 |
| LMArena Japanese | 1393 | 1374 |
| LMArena Korean | 1400 | 1368 |
| LMArena Russian | 1419 | 1417 |
| LMArena Spanish | 1436 | 1405 |
Instruction Following Too close to call
GLM-4.6: 74.3 (#98), GPT-5.4 mini: 74.1 (#102)
| Benchmark | GLM-4.6 | GPT-5.4 mini |
|---|---|---|
| LMArena Instruction Following | 1410 | 1405 |
Long Context Too close to call
GLM-4.6: 43.4 (#94), GPT-5.4 mini: 43.0 (#112)
| Benchmark | GLM-4.6 | GPT-5.4 mini |
|---|---|---|
| LMArena Longer Query | 1422 | 1407 |
Writing & Preference GPT-5.4 mini leads
GLM-4.6: 61.1 (#90), GPT-5.4 mini: 64.0 (#58)
| Benchmark | GLM-4.6 | GPT-5.4 mini |
|---|---|---|
| LMArena Text | 1440 | 1412 |
| LMArena Creative Writing | 1411 | 1370 |
| EQ-Bench Creative Writing | 1411 | 1665 |
| LMArena Multi-Turn | 1427 | 1429 |
Frequently asked questions
Is GLM-4.6 better than GPT-5.4 mini?
GPT-5.4 mini is the stronger model overall, scoring 45.0 to 41.4 on the Noometry Index. GLM-4.6 costs 1.7× less per token, which makes it the better buy when GPT-5.4 mini's lead doesn't matter for your workload.
Which is cheaper, GLM-4.6 or GPT-5.4 mini?
GLM-4.6 is cheaper. It lists at $0.60 per million input tokens and $2.20 per million output tokens; GPT-5.4 mini lists at $0.75 and $4.50.
Is GLM-4.6 or GPT-5.4 mini better for coding?
GPT-5.4 mini scores higher on coding benchmarks: 45.2 versus 40.1 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-4.6 and GPT-5.4 mini share?
26 benchmarks have published results for both models. GLM-4.6 has 29 scored results on Noometry and GPT-5.4 mini has 46.