Model comparison
DeepSeek-R1 vs Phi 3 Mini 4k Instruct
DeepSeek-R1 is the stronger model overall, scoring 42.3 to 27.9 on the Noometry Index.
Last verified . 24 shared benchmarks.
DeepSeek
MicrosoftSummary
- They share 24 benchmarks with published results for both. DeepSeek-R1 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 writing & preference, where DeepSeek-R1 leads 61.4 to 27.6.
- The biggest single-benchmark swing is LiveBench Math: 80.7% for DeepSeek-R1 and 15.7% for Phi 3 Mini 4k Instruct.
- Phi 3 Mini 4k Instruct has downloadable open weights; the other is API-only.
Side by side
| DeepSeek-R1 | Phi 3 Mini 4k Instruct | |
|---|---|---|
| Provider | DeepSeek | Microsoft |
| Noometry Index | 42.3 | 27.9 |
| Released | 2025-01-20 | 2024-04-23 |
| Weights | Proprietary | Open |
| Context window | 164K | — |
| Max output | 64K | — |
| Input $ / M tokens | $0.50 | — |
| Output $ / M tokens | $2.15 | — |
| Results tracked | 52 | 35 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding DeepSeek-R1 leads
DeepSeek-R1: 46.3 (#68), Phi 3 Mini 4k Instruct: 26.6 (#323)
| Benchmark | DeepSeek-R1 | Phi 3 Mini 4k Instruct |
|---|---|---|
| LiveBench Coding | 66.7% | 15.5% |
| LMArena Coding | 1427 | 1093 |
| Aider Polyglot | 71.4% | — |
| SciCode | 35.7% | — |
| WeirdML | 41.6% | — |
| ALE-Bench | 804.12 | — |
| AlgoTune | 1.7 | — |
| HumanEval+ | — | 59.1% |
| MBPP+ | — | 54.2% |
Agentic & Tool Use Not comparable
DeepSeek-R1: 30.7 (#75), Phi 3 Mini 4k Instruct: —
| Benchmark | DeepSeek-R1 | Phi 3 Mini 4k Instruct |
|---|---|---|
| DeepResearch Bench | 35.1% | — |
| BALROG | 34.9% | — |
| METR Time Horizons | 53.8% | — |
Reasoning DeepSeek-R1 leads
DeepSeek-R1: 18.6 (#278), Phi 3 Mini 4k Instruct: 14.1 (#328)
| Benchmark | DeepSeek-R1 | Phi 3 Mini 4k Instruct |
|---|---|---|
| LiveBench Reasoning | 83.2% | 26.8% |
| LMArena Hard Prompts | 1416 | 1072 |
| LiveBench Data Analysis | 69.8% | 34.7% |
| LiveBench | 71.6% | 22.4% |
| ARC-AGI-2 | 1.3% | — |
| SimpleBench | 40.8% | — |
| Kagi LLM Benchmark | 69.4% | — |
| ARC-AGI-1 | 21.2% | — |
| CritPt | 1.1% | — |
| Chess Puzzles | — | 0% |
| Adversarial NLI | — | 52.8% |
| BIG-Bench Hard | — | 71.7% |
| Epoch Capabilities Index | 141.29 | — |
| ForecastBench | 60 | — |
| HellaSwag | — | 76.7% |
| WinoGrande | — | 70.8% |
Math DeepSeek-R1 leads
DeepSeek-R1: 43.8 (#79), Phi 3 Mini 4k Instruct: 26.6 (#257)
| Benchmark | DeepSeek-R1 | Phi 3 Mini 4k Instruct |
|---|---|---|
| LiveBench Math | 80.7% | 15.7% |
| LMArena Math | 1400 | 1111 |
| OTIS Mock AIME 2024-2025 | 66.4% | — |
| Omni-MATH | 42.4% | — |
| MATH Level 5 | 96.6% | — |
Knowledge DeepSeek-R1 leads
DeepSeek-R1: 44.5 (#87), Phi 3 Mini 4k Instruct: 28.5 (#246)
| Benchmark | DeepSeek-R1 | Phi 3 Mini 4k Instruct |
|---|---|---|
| LMArena Expert | 1394 | 1045 |
| GPQA Diamond | 76.3% | — |
| MMLU-Pro | 79.3% | — |
| Confabulations | 12.7% | — |
| Vectara Hallucination Rate | 11.3% | — |
| GPQA (HELM) | 66.6% | — |
| ARC (AI2) Challenge | — | 84.9% |
| MMLU | — | 68.8% |
| OpenBookQA | — | 88% |
| TriviaQA | — | 64% |
Multilingual DeepSeek-R1 leads
DeepSeek-R1: 52.4 (#85), Phi 3 Mini 4k Instruct: 26.3 (#280)
| Benchmark | DeepSeek-R1 | Phi 3 Mini 4k Instruct |
|---|---|---|
| LMArena Non-English | 1412 | 1021 |
| LMArena Chinese | 1442 | 1021 |
| LMArena French | 1417 | 1076 |
| LMArena German | 1404 | 1044 |
| LMArena Japanese | 1391 | 935 |
| LMArena Korean | 1360 | 905 |
| LMArena Russian | 1423 | 1022 |
| LMArena Spanish | 1411 | 1085 |
Instruction Following DeepSeek-R1 leads
DeepSeek-R1: 72.0 (#143), Phi 3 Mini 4k Instruct: 47.7 (#303)
| Benchmark | DeepSeek-R1 | Phi 3 Mini 4k Instruct |
|---|---|---|
| LiveBench Instruction Following | 80.5% | 39.1% |
| LMArena Instruction Following | 1382 | 1053 |
| IFEval | 78.4% | — |
Long Context DeepSeek-R1 leads
DeepSeek-R1: 45.4 (#36), Phi 3 Mini 4k Instruct: 31.7 (#276)
| Benchmark | DeepSeek-R1 | Phi 3 Mini 4k Instruct |
|---|---|---|
| LMArena Longer Query | 1391 | 1044 |
| Fiction.LiveBench | 75% | — |
Writing & Preference DeepSeek-R1 leads
DeepSeek-R1: 61.4 (#88), Phi 3 Mini 4k Instruct: 27.6 (#300)
| Benchmark | DeepSeek-R1 | Phi 3 Mini 4k Instruct |
|---|---|---|
| LMArena Text | 1428 | 1073 |
| LMArena Creative Writing | 1405 | 1037 |
| LMArena Multi-Turn | 1405 | 1018 |
| LiveBench Language | 48.5% | 9.2% |
| Short-Story Creative Writing | 83% | — |
| EQ-Bench Creative Writing | 1500 | — |
| WildBench | 82.8% | — |
Frequently asked questions
Is DeepSeek-R1 better than Phi 3 Mini 4k Instruct?
DeepSeek-R1 is the stronger model overall, scoring 42.3 to 27.9 on the Noometry Index.
Is DeepSeek-R1 or Phi 3 Mini 4k Instruct better for coding?
DeepSeek-R1 scores higher on coding benchmarks: 46.3 versus 26.6 in the Noometry coding category.
How many benchmarks do DeepSeek-R1 and Phi 3 Mini 4k Instruct share?
24 benchmarks have published results for both models. DeepSeek-R1 has 52 scored results on Noometry and Phi 3 Mini 4k Instruct has 35.