Model comparison
Claude Sonnet 4 vs Phi-4
Claude Sonnet 4 is the stronger model overall, scoring 40.8 to 31.2 on the Noometry Index. Phi-4 costs 69× less per token, which makes it the better buy when Claude Sonnet 4's lead doesn't matter for your workload.
Last verified . 24 shared benchmarks.
MicrosoftSummary
- They share 24 benchmarks with published results for both. Claude Sonnet 4 scores higher in 8 categories and Phi-4 in 1 category; 9 gaps are clear of the uncertainty.
- The widest gap is in math, where Claude Sonnet 4 leads 43.3 to 20.8.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 71.1% for Claude Sonnet 4 and 13.8% for Phi-4.
- Phi-4 is cheaper at $0.07 / $0.14 per million input/output tokens, against $3 / $15 for Claude Sonnet 4.
- Claude Sonnet 4 accepts more context: 200K tokens versus 128K.
- Phi-4 has downloadable open weights; the other is API-only.
Side by side
| Claude Sonnet 4 | Phi-4 | |
|---|---|---|
| Provider | Anthropic | Microsoft |
| Noometry Index | 40.8 | 31.2 |
| Released | 2025-05-22 | 2024-12-11 |
| Weights | Proprietary | Open |
| Context window | 200K | 128K |
| Max output | 64K | 4K |
| Input $ / M tokens | $3 | $0.07 |
| Output $ / M tokens | $15 | $0.14 |
| Results tracked | 58 | 37 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Claude Sonnet 4 leads
Claude Sonnet 4: 43.5 (#88), Phi-4: 34.4 (#239)
| Benchmark | Claude Sonnet 4 | Phi-4 |
|---|---|---|
| LMArena Coding | 1414 | 1231 |
| SWE-bench Verified (bash only) | 64.9% | — |
| Aider Polyglot | 61.3% | — |
| SciCode | 40% | — |
| GSO | 4.9% | — |
| WeirdML | 46.1% | — |
| BigCodeBench Instruct | — | 45.5% |
| LiveBench Coding | — | 30.7% |
| BigCodeBench Complete | — | 55.4% |
| ALE-Bench | 655.35 | — |
Agentic & Tool Use Claude Sonnet 4 leads
Claude Sonnet 4: 38.5 (#31), Phi-4: 22.8 (#128)
| Benchmark | Claude Sonnet 4 | Phi-4 |
|---|---|---|
| Berkeley Function Calling Leaderboard | — | 28.8% |
| TheAgentCompany | 33.1% | — |
| Cybench | 35% | — |
| DeepResearch Bench | 46.6% | — |
| OSWorld | 43.9% | — |
| BALROG | — | 11.6% |
| METR Time Horizons | 62% | — |
Reasoning Claude Sonnet 4 leads
Claude Sonnet 4: 22.9 (#187), Phi-4: 17.7 (#291)
| Benchmark | Claude Sonnet 4 | Phi-4 |
|---|---|---|
| LMArena Hard Prompts | 1372 | 1220 |
| Epoch Capabilities Index | 141.69 | 130.42 |
| ARC-AGI-2 | 5.9% | — |
| SimpleBench | 45.5% | — |
| Kagi LLM Benchmark | 73% | — |
| ARC-AGI-1 | 40% | — |
| CritPt | 0.3% | — |
| Chess Puzzles | — | 1% |
| EnigmaEval | 3.1% | — |
| LiveBench Reasoning | — | 47.8% |
| DTBench | 77.1% | — |
| LiveBench Data Analysis | — | 45.2% |
| LMCA | 29% | — |
| ForecastBench | 60.2 | — |
| LiveBench | — | 41.6% |
Math Claude Sonnet 4 leads
Claude Sonnet 4: 43.3 (#80), Phi-4: 20.8 (#285)
| Benchmark | Claude Sonnet 4 | Phi-4 |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 71.1% | 13.8% |
| LMArena Math | 1375 | 1246 |
| MATH Level 5 | 84.4% | 64.9% |
| Omni-MATH | 60.2% | — |
| LiveBench Math | — | 42% |
| FrontierMath (Feb 2025 set) | 4.1% | — |
| FrontierMath Tier 4 (v1) | 0% | — |
Knowledge Claude Sonnet 4 leads
Claude Sonnet 4: 41.8 (#108), Phi-4: 32.6 (#209)
| Benchmark | Claude Sonnet 4 | Phi-4 |
|---|---|---|
| GPQA Diamond | 79.2% | 56.1% |
| Confabulations | 13.2% | 29.4% |
| Vectara Hallucination Rate | 10.3% | 3.7% |
| LMArena Expert | 1372 | 1203 |
| Humanity's Last Exam | 7.8% | — |
| MMLU-Pro | 84.3% | — |
| GPQA (HELM) | 70.6% | — |
| MMLU | — | 84.8% |
Multimodal Not comparable
Claude Sonnet 4: 26.2 (#121), Phi-4: —
| Benchmark | Claude Sonnet 4 | Phi-4 |
|---|---|---|
| LMArena Vision | 1191 | — |
| GeoBench | 37% | — |
| VPCT | 34% | — |
| MindCube | 44.8% | — |
Multilingual Claude Sonnet 4 leads
Claude Sonnet 4: 46.7 (#156), Phi-4: 37.2 (#237)
| Benchmark | Claude Sonnet 4 | Phi-4 |
|---|---|---|
| LMArena Non-English | 1333 | 1197 |
| LMArena Chinese | 1350 | 1212 |
| LMArena French | 1363 | 1224 |
| LMArena German | 1331 | 1222 |
| LMArena Japanese | 1302 | 1158 |
| LMArena Korean | 1291 | 1151 |
| LMArena Russian | 1355 | 1209 |
| LMArena Spanish | 1357 | 1234 |
Instruction Following Claude Sonnet 4 leads
Claude Sonnet 4: 71.7 (#145), Phi-4: 60.4 (#251)
| Benchmark | Claude Sonnet 4 | Phi-4 |
|---|---|---|
| LMArena Instruction Following | 1376 | 1201 |
| LiveBench Instruction Following | — | 58.4% |
| IFEval | 84% | — |
Long Context Phi-4 leads
Claude Sonnet 4: 33.7 (#259), Phi-4: 36.9 (#226)
| Benchmark | Claude Sonnet 4 | Phi-4 |
|---|---|---|
| LMArena Longer Query | 1398 | 1217 |
| Fiction.LiveBench | 46.9% | — |
Writing & Preference Claude Sonnet 4 leads
Claude Sonnet 4: 57.1 (#132), Phi-4: 40.5 (#244)
| Benchmark | Claude Sonnet 4 | Phi-4 |
|---|---|---|
| LMArena Text | 1351 | 1217 |
| LMArena Creative Writing | 1345 | 1182 |
| Short-Story Creative Writing | 81.4% | 62.6% |
| LMArena Multi-Turn | 1376 | 1206 |
| EQ-Bench Creative Writing | 1483 | — |
| WildBench | 83.8% | — |
| LiveBench Language | — | 25.6% |
Frequently asked questions
Is Claude Sonnet 4 better than Phi-4?
Claude Sonnet 4 is the stronger model overall, scoring 40.8 to 31.2 on the Noometry Index. Phi-4 costs 69× less per token, which makes it the better buy when Claude Sonnet 4's lead doesn't matter for your workload.
Which is cheaper, Claude Sonnet 4 or Phi-4?
Phi-4 is cheaper. It lists at $0.07 per million input tokens and $0.14 per million output tokens; Claude Sonnet 4 lists at $3 and $15.
Is Claude Sonnet 4 or Phi-4 better for coding?
Claude Sonnet 4 scores higher on coding benchmarks: 43.5 versus 34.4 in the Noometry coding category.
Which has the bigger context window?
Claude Sonnet 4 does, with 200K tokens against 128K.
How many benchmarks do Claude Sonnet 4 and Phi-4 share?
24 benchmarks have published results for both models. Claude Sonnet 4 has 58 scored results on Noometry and Phi-4 has 37.