Model comparison
Phi-4 vs Qwen3-Coder 480B-A35B Instruct
Qwen3-Coder 480B-A35B Instruct is the stronger model overall, scoring 38.1 to 31.2 on the Noometry Index. Phi-4 costs 34× less per token, which makes it the better buy when Qwen3-Coder 480B-A35B Instruct's lead doesn't matter for your workload.
Last verified . 17 shared benchmarks.
Microsoft
Alibaba (Qwen)Summary
- They share 17 benchmarks with published results for both. Phi-4 scores higher in 0 categories and Qwen3-Coder 480B-A35B Instruct in 9 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in math, where Qwen3-Coder 480B-A35B Instruct leads 37.6 to 20.8.
- Phi-4 is cheaper at $0.07 / $0.14 per million input/output tokens, against $1.50 / $7.50 for Qwen3-Coder 480B-A35B Instruct.
- Qwen3-Coder 480B-A35B Instruct accepts more context: 262K tokens versus 128K.
Side by side
| Phi-4 | Qwen3-Coder 480B-A35B Instruct | |
|---|---|---|
| Provider | Microsoft | Alibaba (Qwen) |
| Noometry Index | 31.2 | 38.1 |
| Released | 2024-12-11 | 2025-04 |
| Weights | Open | Open |
| Context window | 128K | 262K |
| Max output | 4K | 66K |
| Input $ / M tokens | $0.07 | $1.50 |
| Output $ / M tokens | $0.14 | $7.50 |
| Results tracked | 37 | 25 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Qwen3-Coder 480B-A35B Instruct leads
Phi-4: 34.4 (#239), Qwen3-Coder 480B-A35B Instruct: 35.5 (#223)
| Benchmark | Phi-4 | Qwen3-Coder 480B-A35B Instruct |
|---|---|---|
| LMArena Coding | 1231 | 1412 |
| SWE-bench Verified (bash only) | — | 55.4% |
| LMArena WebDev | — | 1275 |
| GSO | — | 4.9% |
| WeirdML | — | 41.2% |
| BigCodeBench Instruct | 45.5% | — |
| LiveBench Coding | 30.7% | — |
| BigCodeBench Complete | 55.4% | — |
| ALE-Bench | — | 461.45 |
| AlgoTune | — | 1.44 |
Agentic & Tool Use Qwen3-Coder 480B-A35B Instruct leads
Phi-4: 22.8 (#128), Qwen3-Coder 480B-A35B Instruct: 23.9 (#123)
| Benchmark | Phi-4 | Qwen3-Coder 480B-A35B Instruct |
|---|---|---|
| Terminal-Bench | — | 27.2% |
| Berkeley Function Calling Leaderboard | 28.8% | — |
| BALROG | 11.6% | — |
Reasoning Qwen3-Coder 480B-A35B Instruct leads
Phi-4: 17.7 (#291), Qwen3-Coder 480B-A35B Instruct: 25.5 (#149)
| Benchmark | Phi-4 | Qwen3-Coder 480B-A35B Instruct |
|---|---|---|
| LMArena Hard Prompts | 1220 | 1372 |
| Kagi LLM Benchmark | — | 49.5% |
| Chess Puzzles | 1% | — |
| LiveBench Reasoning | 47.8% | — |
| LiveBench Data Analysis | 45.2% | — |
| Epoch Capabilities Index | 130.42 | — |
| LiveBench | 41.6% | — |
Math Qwen3-Coder 480B-A35B Instruct leads
Phi-4: 20.8 (#285), Qwen3-Coder 480B-A35B Instruct: 37.6 (#150)
| Benchmark | Phi-4 | Qwen3-Coder 480B-A35B Instruct |
|---|---|---|
| LMArena Math | 1246 | 1365 |
| OTIS Mock AIME 2024-2025 | 13.8% | — |
| LiveBench Math | 42% | — |
| MATH Level 5 | 64.9% | — |
Knowledge Qwen3-Coder 480B-A35B Instruct leads
Phi-4: 32.6 (#209), Qwen3-Coder 480B-A35B Instruct: 37.0 (#162)
| Benchmark | Phi-4 | Qwen3-Coder 480B-A35B Instruct |
|---|---|---|
| LMArena Expert | 1203 | 1338 |
| GPQA Diamond | 56.1% | — |
| Confabulations | 29.4% | — |
| Vectara Hallucination Rate | 3.7% | — |
| MMLU | 84.8% | — |
Multilingual Qwen3-Coder 480B-A35B Instruct leads
Phi-4: 37.2 (#237), Qwen3-Coder 480B-A35B Instruct: 47.7 (#148)
| Benchmark | Phi-4 | Qwen3-Coder 480B-A35B Instruct |
|---|---|---|
| LMArena Non-English | 1197 | 1346 |
| LMArena Chinese | 1212 | 1357 |
| LMArena French | 1224 | 1398 |
| LMArena German | 1222 | 1325 |
| LMArena Japanese | 1158 | 1310 |
| LMArena Korean | 1151 | 1305 |
| LMArena Russian | 1209 | 1366 |
| LMArena Spanish | 1234 | 1360 |
Instruction Following Qwen3-Coder 480B-A35B Instruct leads
Phi-4: 60.4 (#251), Qwen3-Coder 480B-A35B Instruct: 71.6 (#147)
| Benchmark | Phi-4 | Qwen3-Coder 480B-A35B Instruct |
|---|---|---|
| LMArena Instruction Following | 1201 | 1355 |
| LiveBench Instruction Following | 58.4% | — |
Long Context Qwen3-Coder 480B-A35B Instruct leads
Phi-4: 36.9 (#226), Qwen3-Coder 480B-A35B Instruct: 42.0 (#131)
| Benchmark | Phi-4 | Qwen3-Coder 480B-A35B Instruct |
|---|---|---|
| LMArena Longer Query | 1217 | 1378 |
Writing & Preference Qwen3-Coder 480B-A35B Instruct leads
Phi-4: 40.5 (#244), Qwen3-Coder 480B-A35B Instruct: 55.3 (#147)
| Benchmark | Phi-4 | Qwen3-Coder 480B-A35B Instruct |
|---|---|---|
| LMArena Text | 1217 | 1357 |
| LMArena Creative Writing | 1182 | 1333 |
| LMArena Multi-Turn | 1206 | 1365 |
| Short-Story Creative Writing | 62.6% | — |
| LiveBench Language | 25.6% | — |
Frequently asked questions
Is Phi-4 better than Qwen3-Coder 480B-A35B Instruct?
Qwen3-Coder 480B-A35B Instruct is the stronger model overall, scoring 38.1 to 31.2 on the Noometry Index. Phi-4 costs 34× less per token, which makes it the better buy when Qwen3-Coder 480B-A35B Instruct's lead doesn't matter for your workload.
Which is cheaper, Phi-4 or Qwen3-Coder 480B-A35B Instruct?
Phi-4 is cheaper. It lists at $0.07 per million input tokens and $0.14 per million output tokens; Qwen3-Coder 480B-A35B Instruct lists at $1.50 and $7.50.
Is Phi-4 or Qwen3-Coder 480B-A35B Instruct better for coding?
Qwen3-Coder 480B-A35B Instruct scores higher on coding benchmarks: 35.5 versus 34.4 in the Noometry coding category.
Which has the bigger context window?
Qwen3-Coder 480B-A35B Instruct does, with 262K tokens against 128K.
How many benchmarks do Phi-4 and Qwen3-Coder 480B-A35B Instruct share?
17 benchmarks have published results for both models. Phi-4 has 37 scored results on Noometry and Qwen3-Coder 480B-A35B Instruct has 25.