Model comparison
DeepSeek V4 Pro vs Phi 3 Mini 4k Instruct
DeepSeek V4 Pro is the stronger model overall, scoring 54.3 to 27.9 on the Noometry Index.
Last verified . 18 shared benchmarks.
DeepSeek
MicrosoftSummary
- They share 18 benchmarks with published results for both. DeepSeek V4 Pro scores higher in 8 categories and Phi 3 Mini 4k Instruct in 0 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where DeepSeek V4 Pro leads 56.5 to 14.1.
- The biggest single-benchmark swing is Chess Puzzles: 47% for DeepSeek V4 Pro and 0% for Phi 3 Mini 4k Instruct.
Side by side
| DeepSeek V4 Pro | Phi 3 Mini 4k Instruct | |
|---|---|---|
| Provider | DeepSeek | Microsoft |
| Noometry Index | 54.3 | 27.9 |
| Released | 2026-04-24 | 2024-04-23 |
| Weights | Open | Open |
| Context window | 1M | — |
| Max output | 393K | — |
| Input $ / M tokens | $0.66 | — |
| Output $ / M tokens | $1.98 | — |
| Results tracked | 48 | 35 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding DeepSeek V4 Pro leads
DeepSeek V4 Pro: 52.4 (#34), Phi 3 Mini 4k Instruct: 26.6 (#323)
| Benchmark | DeepSeek V4 Pro | Phi 3 Mini 4k Instruct |
|---|---|---|
| LMArena Coding | 1470 | 1093 |
| SWE-bench Verified | 77.6% | — |
| FrontierCode | 28.6% | — |
| LMArena WebDev | 1582 | — |
| SciCode | 51% | — |
| WeirdML | 66.2% | — |
| LiveBench Coding | — | 15.5% |
| ALE-Bench | 1,403 | — |
| HumanEval+ | — | 59.1% |
| MBPP+ | — | 54.2% |
Agentic & Tool Use Not comparable
DeepSeek V4 Pro: 32.8 (#58), Phi 3 Mini 4k Instruct: —
| Benchmark | DeepSeek V4 Pro | Phi 3 Mini 4k Instruct |
|---|---|---|
| APEX-Agents | 47.3% | — |
| Vending-Bench 2 | 3,285 | — |
Reasoning DeepSeek V4 Pro leads
DeepSeek V4 Pro: 56.5 (#24), Phi 3 Mini 4k Instruct: 14.1 (#328)
| Benchmark | DeepSeek V4 Pro | Phi 3 Mini 4k Instruct |
|---|---|---|
| Chess Puzzles | 47% | 0% |
| LMArena Hard Prompts | 1461 | 1072 |
| ARC-AGI-2 | 61.3% | — |
| Kagi LLM Benchmark | 53.5% | — |
| NYT Connections (extended) | 91.3% | — |
| ARC-AGI-1 | 90.5% | — |
| CritPt | 18% | — |
| LiveBench Reasoning | — | 26.8% |
| Mystery Game Puzzles | 43% | — |
| DTBench | 93.9% | — |
| LiveBench Data Analysis | — | 34.7% |
| LMCA | 45.5% | — |
| Surface Evolver Bench | 40% | — |
| Adversarial NLI | — | 52.8% |
| BIG-Bench Hard | — | 71.7% |
| Epoch Capabilities Index | 155.31 | — |
| ForecastBench | 56.1 | — |
| HellaSwag | — | 76.7% |
| LiveBench | — | 22.4% |
| WinoGrande | — | 70.8% |
Math DeepSeek V4 Pro leads
DeepSeek V4 Pro: 64.8 (#30), Phi 3 Mini 4k Instruct: 26.6 (#257)
| Benchmark | DeepSeek V4 Pro | Phi 3 Mini 4k Instruct |
|---|---|---|
| LMArena Math | 1455 | 1111 |
| FrontierMath (Tiers 1-3) | 64.6% | — |
| FrontierMath Tier 4 | 26.8% | — |
| MathArena Final-Answer Competitions | 76.6% | — |
| OTIS Mock AIME 2024-2025 | 98.6% | — |
| ProofBench | 50% | — |
| LiveBench Math | — | 15.7% |
Knowledge DeepSeek V4 Pro leads
DeepSeek V4 Pro: 59.5 (#31), Phi 3 Mini 4k Instruct: 28.5 (#246)
| Benchmark | DeepSeek V4 Pro | Phi 3 Mini 4k Instruct |
|---|---|---|
| LMArena Expert | 1464 | 1045 |
| GPQA Diamond | 91.7% | — |
| SimpleQA Verified | 52.9% | — |
| Vectara Hallucination Rate | 8.6% | — |
| ARC (AI2) Challenge | — | 84.9% |
| MMLU | — | 68.8% |
| OpenBookQA | — | 88% |
| TriviaQA | — | 64% |
Multilingual DeepSeek V4 Pro leads
DeepSeek V4 Pro: 54.4 (#45), Phi 3 Mini 4k Instruct: 26.3 (#280)
| Benchmark | DeepSeek V4 Pro | Phi 3 Mini 4k Instruct |
|---|---|---|
| LMArena Non-English | 1439 | 1021 |
| LMArena Chinese | 1486 | 1021 |
| LMArena French | 1472 | 1076 |
| LMArena German | 1458 | 1044 |
| LMArena Japanese | 1445 | 935 |
| LMArena Korean | 1447 | 905 |
| LMArena Russian | 1453 | 1022 |
| LMArena Spanish | 1458 | 1085 |
Instruction Following DeepSeek V4 Pro leads
DeepSeek V4 Pro: 76.1 (#47), Phi 3 Mini 4k Instruct: 47.7 (#303)
| Benchmark | DeepSeek V4 Pro | Phi 3 Mini 4k Instruct |
|---|---|---|
| LMArena Instruction Following | 1448 | 1053 |
| LiveBench Instruction Following | — | 39.1% |
Long Context DeepSeek V4 Pro leads
DeepSeek V4 Pro: 45.0 (#51), Phi 3 Mini 4k Instruct: 31.7 (#276)
| Benchmark | DeepSeek V4 Pro | Phi 3 Mini 4k Instruct |
|---|---|---|
| LMArena Longer Query | 1458 | 1044 |
| CL-bench Life | 13.5% | — |
Writing & Preference DeepSeek V4 Pro leads
DeepSeek V4 Pro: 65.5 (#46), Phi 3 Mini 4k Instruct: 27.6 (#300)
| Benchmark | DeepSeek V4 Pro | Phi 3 Mini 4k Instruct |
|---|---|---|
| LMArena Text | 1451 | 1073 |
| LMArena Creative Writing | 1446 | 1037 |
| LMArena Multi-Turn | 1467 | 1018 |
| EQ-Bench Creative Writing | 1553 | — |
| EQ-Bench 4 | 1166 | — |
| LiveBench Language | — | 9.2% |
Frequently asked questions
Is DeepSeek V4 Pro better than Phi 3 Mini 4k Instruct?
DeepSeek V4 Pro is the stronger model overall, scoring 54.3 to 27.9 on the Noometry Index.
Is DeepSeek V4 Pro or Phi 3 Mini 4k Instruct better for coding?
DeepSeek V4 Pro scores higher on coding benchmarks: 52.4 versus 26.6 in the Noometry coding category.
How many benchmarks do DeepSeek V4 Pro and Phi 3 Mini 4k Instruct share?
18 benchmarks have published results for both models. DeepSeek V4 Pro has 48 scored results on Noometry and Phi 3 Mini 4k Instruct has 35.