Model comparison
GLM-5 vs Phi-4
GLM-5 is the stronger model overall, scoring 46.1 to 31.2 on the Noometry Index. Phi-4 costs 18× less per token, which makes it the better buy when GLM-5's lead doesn't matter for your workload.
Last verified . 22 shared benchmarks.
MicrosoftSummary
- They share 22 benchmarks with published results for both. GLM-5 scores higher in 9 categories and Phi-4 in 0 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in math, where GLM-5 leads 46.4 to 20.8.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 80% for GLM-5 and 13.8% for Phi-4.
- Phi-4 is cheaper at $0.07 / $0.14 per million input/output tokens, against $1 / $3.20 for GLM-5.
- GLM-5 accepts more context: 205K tokens versus 128K.
Side by side
| GLM-5 | Phi-4 | |
|---|---|---|
| Provider | Z.ai (Zhipu) | Microsoft |
| Noometry Index | 46.1 | 31.2 |
| Released | 2026-02-11 | 2024-12-11 |
| Weights | Open | Open |
| Context window | 205K | 128K |
| Max output | 131K | 4K |
| Input $ / M tokens | $1 | $0.07 |
| Output $ / M tokens | $3.20 | $0.14 |
| Results tracked | 45 | 37 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding GLM-5 leads
GLM-5: 49.0 (#52), Phi-4: 34.4 (#239)
| Benchmark | GLM-5 | Phi-4 |
|---|---|---|
| LMArena Coding | 1461 | 1231 |
| SWE-bench Verified | 72.1% | — |
| SWE-bench Verified (bash only) | 72.8% | — |
| LMArena WebDev | 1434 | — |
| SWE-bench Multilingual | 69.7% | — |
| WeirdML | 48.2% | — |
| BigCodeBench Instruct | — | 45.5% |
| LiveBench Coding | — | 30.7% |
| BigCodeBench Complete | — | 55.4% |
| ALE-Bench | 765.62 | — |
Agentic & Tool Use GLM-5 leads
GLM-5: 31.1 (#71), Phi-4: 22.8 (#128)
| Benchmark | GLM-5 | Phi-4 |
|---|---|---|
| Terminal-Bench | 52.4% | — |
| Berkeley Function Calling Leaderboard | — | 28.8% |
| τ²-bench Airline | 82.5% | — |
| τ²-bench Banking | 9.8% | — |
| τ²-bench Retail | 73.7% | — |
| τ²-bench Telecom | 86.8% | — |
| BALROG | — | 11.6% |
| Vending-Bench 2 | 4,432 | — |
Reasoning GLM-5 leads
GLM-5: 27.6 (#116), Phi-4: 17.7 (#291)
| Benchmark | GLM-5 | Phi-4 |
|---|---|---|
| Chess Puzzles | 10% | 1% |
| LMArena Hard Prompts | 1452 | 1220 |
| Epoch Capabilities Index | 145.83 | 130.42 |
| ARC-AGI-2 | 4.9% | — |
| SimpleBench | 53.2% | — |
| Kagi LLM Benchmark | 75% | — |
| NYT Connections (extended) | 74.8% | — |
| ARC-AGI-1 | 44.7% | — |
| LiveBench Reasoning | — | 47.8% |
| LiveBench Data Analysis | — | 45.2% |
| ForecastBench | 61 | — |
| LiveBench | — | 41.6% |
Math GLM-5 leads
GLM-5: 46.4 (#71), Phi-4: 20.8 (#285)
| Benchmark | GLM-5 | Phi-4 |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 80% | 13.8% |
| LMArena Math | 1440 | 1246 |
| MathArena Final-Answer Competitions | 65.7% | — |
| LiveBench Math | — | 42% |
| MATH Level 5 | — | 64.9% |
| FrontierMath (Feb 2025 set) | 16.4% | — |
| FrontierMath Tier 4 (v1) | 2.1% | — |
Knowledge GLM-5 leads
GLM-5: 52.3 (#64), Phi-4: 32.6 (#209)
| Benchmark | GLM-5 | Phi-4 |
|---|---|---|
| GPQA Diamond | 87.8% | 56.1% |
| Vectara Hallucination Rate | 10.1% | 3.7% |
| LMArena Expert | 1454 | 1203 |
| Confabulations | — | 29.4% |
| MMLU | — | 84.8% |
Multilingual GLM-5 leads
GLM-5: 53.7 (#58), Phi-4: 37.2 (#237)
| Benchmark | GLM-5 | Phi-4 |
|---|---|---|
| LMArena Non-English | 1430 | 1197 |
| LMArena Chinese | 1511 | 1212 |
| LMArena French | 1455 | 1224 |
| LMArena German | 1445 | 1222 |
| LMArena Japanese | 1416 | 1158 |
| LMArena Korean | 1423 | 1151 |
| LMArena Russian | 1436 | 1209 |
| LMArena Spanish | 1454 | 1234 |
Instruction Following GLM-5 leads
GLM-5: 75.2 (#67), Phi-4: 60.4 (#251)
| Benchmark | GLM-5 | Phi-4 |
|---|---|---|
| LMArena Instruction Following | 1428 | 1201 |
| LiveBench Instruction Following | — | 58.4% |
Long Context GLM-5 leads
GLM-5: 44.7 (#60), Phi-4: 36.9 (#226)
| Benchmark | GLM-5 | Phi-4 |
|---|---|---|
| LMArena Longer Query | 1446 | 1217 |
| CL-bench | 18.7% | — |
Writing & Preference GLM-5 leads
GLM-5: 66.0 (#38), Phi-4: 40.5 (#244)
| Benchmark | GLM-5 | Phi-4 |
|---|---|---|
| LMArena Text | 1446 | 1217 |
| LMArena Creative Writing | 1439 | 1182 |
| LMArena Multi-Turn | 1456 | 1206 |
| Short-Story Creative Writing | — | 62.6% |
| EQ-Bench Creative Writing | 1601 | — |
| LiveBench Language | — | 25.6% |
Frequently asked questions
Is GLM-5 better than Phi-4?
GLM-5 is the stronger model overall, scoring 46.1 to 31.2 on the Noometry Index. Phi-4 costs 18× 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 Phi-4?
Phi-4 is cheaper. It lists at $0.07 per million input tokens and $0.14 per million output tokens; GLM-5 lists at $1 and $3.20.
Is GLM-5 or Phi-4 better for coding?
GLM-5 scores higher on coding benchmarks: 49.0 versus 34.4 in the Noometry coding category.
Which has the bigger context window?
GLM-5 does, with 205K tokens against 128K.
How many benchmarks do GLM-5 and Phi-4 share?
22 benchmarks have published results for both models. GLM-5 has 45 scored results on Noometry and Phi-4 has 37.