Compare / Qwen3-32B vs Phi-4
Qwen3-32B vs Phi-4
Which is cheaper, which is more capable, and which does better at each kind of work, from 26 results on 5 sources that measured both models.
Of the 26 results both models have, the better value on each. Results whose reported ranges overlap are too close to call.
Results as of 9 October 2026 (catalogue release 2026-10-09-dd620e778ac5).
Qwen3-32B or Phi-4? The short answer
Pick Qwen3-32B if quality matters most: it scores higher (Intelligence Index 8.6 against 5.9). Pick Phi-4 if cost matters more: it costs 2.3 times less per token.
- Phi-4 is cheaper: $0.088 against $0.21 per million tokens, 2.3 times less.
- Qwen3-32B scores higher on the Artificial Analysis Intelligence Index: 8.6 against 5.9.
- Qwen3-32B is newer: released 28 Apr 2025, against 12 Dec 2024 for Phi-4.
- On third-party benchmarks, on the results both models have: Qwen3-32B does better for agents and tool use, reasoning and knowledge and writing; Phi-4 for sticking to the facts.
Head to head
Every fact side by side
| Qwen3-32B | Phi-4 | |
|---|---|---|
| Price per million tokens | $0.14 in · $0.40 out | $0.070 in · $0.14 out |
| Blended (3 in : 1 out) | $0.21 | $0.088 |
| Cheapest host, blended | $0.13 (DeepInfra, FP8) | – |
| Intelligence Index | 8.6 | 5.9 |
| Spring Prompt overall | – | – |
| Output speed, tokens a second | – | 44 |
| Context window | 131,072 tokens | 16,384 tokens |
| Released | 28 Apr 2025 (newer) | 12 Dec 2024 |
| Weights | Open | Open |
| Developer based in | China | the US |
Prices are the developer's list price where we have read it, otherwise the typical host on OpenRouter; see every model's price. Intelligence and speed from Artificial Analysis.
Which is better for what
Each model's rank among the models on sale today, for each kind of work, from our own benchmarks and licensed sources. Further right is better; the better rank is in green.
Every shared result
Each model's best result on every measure both have, grouped by benchmark, with the setting that produced it. Under each measure, every model that benchmark has measured is a grey tick, best to the right, so you can see where the two sit in the field. The headline measures of each benchmark come first; the rest are under "Show all". Green marks the better value; where a source reports ranges that overlap, neither is marked.
| Measure | Qwen3-32B | Phi-4 |
|---|---|---|
| Arena (formerly LMArena) reported by Arena (formerly LMArena) · 3 results | ||
| Overall · TextArena rating, higher is better 195 · 291 of 387 measured | 1,347 | 1,256 |
| Business, management and finance · TextArena rating, higher is better 203 · 282 of 380 measured | 1,340 | 1,256 |
| Creative writing · TextArena rating, higher is better 205 · 290 of 385 measured | 1,304 | 1,210 |
| Artificial Analysis reported by Artificial Analysis · 3 results | ||
| Artificial Analysis Intelligence Indexindex score, higher is better 188 · 231 of 243 measured | 8.6thinking reasoning | 5.9 |
| GPQA Diamond% of questions, higher is better 167 · 190 of 231 measured | 66.8%thinking reasoning | 57.5% |
| Humanity's Last Exam% of questions, higher is better 168 · 221 of 242 measured | 7.4%thinking reasoning | 3.8% |
| Berkeley Function Calling Leaderboard (BFCL) V4 reported by Berkeley Function Calling Leaderboard (BFCL) V4 · 3 results | ||
| Overall accuracy · BFCL% correct, higher is better 24 · 51 of 81 measured | 48.7%bfcl fc | 28.8%bfcl prompt |
| Irrelevance detection · BFCL% correct, higher is better 37 · 16 of 81 measured | 82.4%bfcl prompt | 87.6%bfcl prompt |
| Memory · BFCL% correct, higher is better 21 · 24 of 81 measured | 26.7%bfcl fc | 24.7%bfcl prompt |
| UGI Leaderboard reported by UGI Leaderboard · 3 results | ||
| Writing score · UGIscore out of 100, higher is better 134 · 175 of 230 measured | 33.0no reasoning | 25.7 |
| Requested-length error · UGI% off the requested word count, lower is better 99 · 99 of 230 measured | 19.0%no reasoning | 19.0% |
| Style adherence · UGIscore from 0 to 1, higher is better 175 · 122 of 230 measured | 0.32no reasoning | 0.34 |
| Vectara Hallucination Leaderboard reported by Vectara Hallucination Leaderboard · 2 results | ||
| Hallucination rate · Vectara% of summaries, lower is better 24 · 4 of 104 measured | 5.9% | 3.7% |
| Answer rate · Vectara% of documents, higher is better 17 · 100 of 104 measured | 99.9% | 80.7% |
Show all 26 results (12 more)
| Measure | Qwen3-32B | Phi-4 |
|---|---|---|
| Arena (formerly LMArena) reported by Arena (formerly LMArena) · 3 results | ||
| Expert prompts · TextArena rating, higher is better 159 · 261 of 338 measured | 1,398 | 1,268 |
| Instruction following · TextArena rating, higher is better 198 · 285 of 387 measured | 1,331 | 1,245 |
| Writing, literature and language · TextArena rating, higher is better 217 · 290 of 386 measured | 1,308 | 1,229 |
| Artificial Analysis reported by Artificial Analysis · 4 results | ||
| IFBench% of instructions, higher is better 155 · 192 of 195 measured | 36.3%thinking reasoning | 23.5% |
| Long-context reasoning (AA-LCR)% of questions, higher is better 222 · 222 of 233 measured | 0.0%no reasoning | 0.0% |
| Terminal-Bench Hard% of tasks, higher is better 169 · 162 of 190 measured | 3.0%thinking reasoning | 3.8% |
| τ²-bench telecom% of tasks, higher is better 135 · 188 of 193 measured | 29.8%thinking reasoning | 0.0% |
| Berkeley Function Calling Leaderboard (BFCL) V4 reported by Berkeley Function Calling Leaderboard (BFCL) V4 · 5 results | ||
| Multi-turn tasks · BFCL% correct, higher is better 17 · 64 of 81 measured | 47.9%bfcl fc | 3.9%bfcl prompt |
| Relevance detection · BFCL% correct, higher is better 8 · 74 of 81 measured | 93.8%bfcl fc | 50.0%bfcl prompt |
| Single-turn calls (curated) · BFCL% correct, higher is better 3 · 69 of 81 measured | 90.3%bfcl prompt | 69.6%bfcl prompt |
| Single-turn calls (user-contributed) · BFCL% correct, higher is better 3 · 66 of 81 measured | 82.0%bfcl fc | 60.7%bfcl prompt |
| Web search · BFCL% correct, higher is better 28 · 44 of 81 measured | 26.0%bfcl prompt | 4.5%bfcl prompt |
Quick answers
Is Qwen3-32B better than Phi-4?
Pick Qwen3-32B if quality matters most: it scores higher (Intelligence Index 8.6 against 5.9). Pick Phi-4 if cost matters more: it costs 2.3 times less per token.
Which is cheaper, Qwen3-32B or Phi-4?
Phi-4 is cheaper: $0.088 against $0.21 per million tokens, 2.3 times less.
Which is better for agents and tool use, Qwen3-32B or Phi-4?
Qwen3-32B ranks 113 of 164 models on sale for multi-step tasks with tools: support desks, coding agents, function calls, against 123 for Phi-4.
Which is better for reasoning and knowledge, Qwen3-32B or Phi-4?
Qwen3-32B ranks 148 of 198 models on sale for hard questions across science, maths and general knowledge, against 181 for Phi-4.
Which is better for writing, Qwen3-32B or Phi-4?
Qwen3-32B ranks 139 of 188 models on sale for what people prefer in blind comparisons, and judged writing quality, against 176 for Phi-4.
Which is better for sticking to the facts, Qwen3-32B or Phi-4?
Phi-4 ranks 3 of 152 models on sale for summarising without inventing things, and factual answers, against 22 for Qwen3-32B.
Which has the bigger context window, Qwen3-32B or Phi-4?
Qwen3-32B: 131,072 tokens, against 16,384 for Phi-4.
Run Qwen3-32B and Phi-4 on your own prompt
Benchmarks aren't your data. Try both side by side in the playground, with the cost of every answer.