Model comparison
Llama 3.1-405B vs Phi-4
Llama 3.1-405B and Phi-4 score almost the same on the Noometry Index (30.7 vs 31.2), so choose on price, context window or the category you care about most.
Last verified . 23 shared benchmarks.
Meta
MicrosoftSummary
- They share 23 benchmarks with published results for both. Llama 3.1-405B scores higher in 3 categories and Phi-4 in 6 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in instruction following, where Llama 3.1-405B leads 65.9 to 60.4.
- The biggest single-benchmark swing is MATH Level 5: 49.8% for Llama 3.1-405B and 64.9% for Phi-4.
Side by side
| Llama 3.1-405B | Phi-4 | |
|---|---|---|
| Provider | Meta | Microsoft |
| Noometry Index | 30.7 | 31.2 |
| Released | 2024-07-23 | 2024-12-11 |
| Weights | Open | Open |
| Context window | — | 128K |
| Max output | — | 4K |
| Input $ / M tokens | — | $0.07 |
| Output $ / M tokens | — | $0.14 |
| Results tracked | 42 | 37 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Phi-4 leads
Llama 3.1-405B: 33.1 (#262), Phi-4: 34.4 (#239)
| Benchmark | Llama 3.1-405B | Phi-4 |
|---|---|---|
| LMArena Coding | 1291 | 1231 |
| WeirdML | 21.4% | — |
| BigCodeBench Instruct | — | 45.5% |
| LiveBench Coding | — | 30.7% |
| BigCodeBench Complete | — | 55.4% |
Agentic & Tool Use Phi-4 leads
Llama 3.1-405B: 21.0 (#140), Phi-4: 22.8 (#128)
| Benchmark | Llama 3.1-405B | Phi-4 |
|---|---|---|
| Berkeley Function Calling Leaderboard | — | 28.8% |
| TheAgentCompany | 7.4% | — |
| Cybench | 7.5% | — |
| BALROG | — | 11.6% |
Reasoning Too close to call
Llama 3.1-405B: 16.8 (#300), Phi-4: 17.7 (#291)
| Benchmark | Llama 3.1-405B | Phi-4 |
|---|---|---|
| LMArena Hard Prompts | 1269 | 1220 |
| Epoch Capabilities Index | 128.75 | 130.42 |
| SimpleBench | 23% | — |
| Kagi LLM Benchmark | 45% | — |
| Chess Puzzles | — | 1% |
| LiveBench Reasoning | — | 47.8% |
| DTBench | 61.4% | — |
| LiveBench Data Analysis | — | 45.2% |
| BIG-Bench Hard | 82.9% | — |
| ForecastBench | 59.9 | — |
| HellaSwag | 89.2% | — |
| LiveBench | — | 41.6% |
| PIQA | 85.9% | — |
| WinoGrande | 89.2% | — |
Math Phi-4 leads
Llama 3.1-405B: 18.4 (#290), Phi-4: 20.8 (#285)
| Benchmark | Llama 3.1-405B | Phi-4 |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 9.7% | 13.8% |
| LMArena Math | 1281 | 1246 |
| MATH Level 5 | 49.8% | 64.9% |
| Omni-MATH | 24.9% | — |
| LiveBench Math | — | 42% |
Knowledge Phi-4 leads
Llama 3.1-405B: 30.4 (#227), Phi-4: 32.6 (#209)
| Benchmark | Llama 3.1-405B | Phi-4 |
|---|---|---|
| GPQA Diamond | 50.9% | 56.1% |
| Confabulations | 17.6% | 29.4% |
| LMArena Expert | 1243 | 1203 |
| MMLU | 84.5% | 84.8% |
| MMLU-Pro | 72.3% | — |
| Vectara Hallucination Rate | — | 3.7% |
| GPQA (HELM) | 52.2% | — |
| ARC (AI2) Challenge | 95.3% | — |
| TriviaQA | 82.7% | — |
Multilingual Llama 3.1-405B leads
Llama 3.1-405B: 40.7 (#214), Phi-4: 37.2 (#237)
| Benchmark | Llama 3.1-405B | Phi-4 |
|---|---|---|
| LMArena Non-English | 1248 | 1197 |
| LMArena Chinese | 1242 | 1212 |
| LMArena French | 1279 | 1224 |
| LMArena German | 1252 | 1222 |
| LMArena Japanese | 1208 | 1158 |
| LMArena Korean | 1184 | 1151 |
| LMArena Russian | 1265 | 1209 |
| LMArena Spanish | 1260 | 1234 |
Instruction Following Llama 3.1-405B leads
Llama 3.1-405B: 65.9 (#214), Phi-4: 60.4 (#251)
| Benchmark | Llama 3.1-405B | Phi-4 |
|---|---|---|
| LMArena Instruction Following | 1259 | 1201 |
| LiveBench Instruction Following | — | 58.4% |
| IFEval | 81.1% | — |
Long Context Llama 3.1-405B leads
Llama 3.1-405B: 38.4 (#197), Phi-4: 36.9 (#226)
| Benchmark | Llama 3.1-405B | Phi-4 |
|---|---|---|
| LMArena Longer Query | 1266 | 1217 |
Writing & Preference Phi-4 leads
Llama 3.1-405B: 38.9 (#251), Phi-4: 40.5 (#244)
| Benchmark | Llama 3.1-405B | Phi-4 |
|---|---|---|
| LMArena Text | 1284 | 1217 |
| LMArena Creative Writing | 1262 | 1182 |
| LMArena Multi-Turn | 1297 | 1206 |
| Short-Story Creative Writing | — | 62.6% |
| EQ-Bench Creative Writing | 870 | — |
| WildBench | 78.3% | — |
| LiveBench Language | — | 25.6% |
Frequently asked questions
Is Llama 3.1-405B better than Phi-4?
Llama 3.1-405B and Phi-4 score almost the same on the Noometry Index (30.7 vs 31.2), so choose on price, context window or the category you care about most.
Is Llama 3.1-405B or Phi-4 better for coding?
Phi-4 scores higher on coding benchmarks: 34.4 versus 33.1 in the Noometry coding category.
How many benchmarks do Llama 3.1-405B and Phi-4 share?
23 benchmarks have published results for both models. Llama 3.1-405B has 42 scored results on Noometry and Phi-4 has 37.