Model comparison
DeepSeek-V3.2-Exp vs Phi 3 Mini 4k Instruct
DeepSeek-V3.2-Exp is the stronger model overall, scoring 44.3 to 27.9 on the Noometry Index.
Last verified . 18 shared benchmarks.
DeepSeek
MicrosoftSummary
- They share 18 benchmarks with published results for both. DeepSeek-V3.2-Exp 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-V3.2-Exp leads 62.4 to 27.6.
- The biggest single-benchmark swing is Chess Puzzles: 14% for DeepSeek-V3.2-Exp and 0% for Phi 3 Mini 4k Instruct.
Side by side
| DeepSeek-V3.2-Exp | Phi 3 Mini 4k Instruct | |
|---|---|---|
| Provider | DeepSeek | Microsoft |
| Noometry Index | 44.3 | 27.9 |
| Released | 2025-09-29 | 2024-04-23 |
| Weights | Open | Open |
| Context window | 164K | — |
| Max output | 66K | — |
| Input $ / M tokens | $0.26 | — |
| Output $ / M tokens | $0.38 | — |
| Results tracked | 49 | 35 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding DeepSeek-V3.2-Exp leads
DeepSeek-V3.2-Exp: 46.5 (#65), Phi 3 Mini 4k Instruct: 26.6 (#323)
| Benchmark | DeepSeek-V3.2-Exp | Phi 3 Mini 4k Instruct |
|---|---|---|
| LMArena Coding | 1454 | 1093 |
| SWE-bench Verified (bash only) | 70% | — |
| Aider Polyglot | 74.2% | — |
| LMArena WebDev | 1362 | — |
| SWE-bench Multilingual | 59% | — |
| SciCode | 38.9% | — |
| WeirdML | 39.5% | — |
| LiveBench Coding | — | 15.5% |
| HumanEval+ | — | 59.1% |
| MBPP+ | — | 54.2% |
Agentic & Tool Use Not comparable
DeepSeek-V3.2-Exp: 32.7 (#59), Phi 3 Mini 4k Instruct: —
| Benchmark | DeepSeek-V3.2-Exp | Phi 3 Mini 4k Instruct |
|---|---|---|
| Terminal-Bench | 39.6% | — |
| APEX-Agents | 21.3% | — |
| Berkeley Function Calling Leaderboard | 56.7% | — |
| TheAgentCompany | 42.9% | — |
| Vending-Bench 2 | 1,034 | — |
Reasoning DeepSeek-V3.2-Exp leads
DeepSeek-V3.2-Exp: 22.1 (#208), Phi 3 Mini 4k Instruct: 14.1 (#328)
| Benchmark | DeepSeek-V3.2-Exp | Phi 3 Mini 4k Instruct |
|---|---|---|
| Chess Puzzles | 14% | 0% |
| LMArena Hard Prompts | 1434 | 1072 |
| ARC-AGI-2 | 4% | — |
| Kagi LLM Benchmark | 52.2% | — |
| NYT Connections (extended) | 36.7% | — |
| ARC-AGI-1 | 57% | — |
| CritPt | 2.9% | — |
| Thematic Generalization | 65% | — |
| LiveBench Reasoning | — | 26.8% |
| DTBench | 87.7% | — |
| LiveBench Data Analysis | — | 34.7% |
| LMCA | 29.1% | — |
| Adversarial NLI | — | 52.8% |
| BIG-Bench Hard | — | 71.7% |
| Epoch Capabilities Index | 146.27 | — |
| HellaSwag | — | 76.7% |
| LiveBench | — | 22.4% |
| WinoGrande | — | 70.8% |
Math DeepSeek-V3.2-Exp leads
DeepSeek-V3.2-Exp: 41.7 (#87), Phi 3 Mini 4k Instruct: 26.6 (#257)
| Benchmark | DeepSeek-V3.2-Exp | Phi 3 Mini 4k Instruct |
|---|---|---|
| LMArena Math | 1435 | 1111 |
| MathArena Final-Answer Competitions | 57.7% | — |
| OTIS Mock AIME 2024-2025 | 87.8% | — |
| ProofBench | 8% | — |
| LiveBench Math | — | 15.7% |
| FrontierMath (Feb 2025 set) | 22.1% | — |
| FrontierMath Tier 4 (v1) | 2.1% | — |
Knowledge DeepSeek-V3.2-Exp leads
DeepSeek-V3.2-Exp: 51.7 (#66), Phi 3 Mini 4k Instruct: 28.5 (#246)
| Benchmark | DeepSeek-V3.2-Exp | Phi 3 Mini 4k Instruct |
|---|---|---|
| LMArena Expert | 1436 | 1045 |
| GPQA Diamond | 83.4% | — |
| Vectara Hallucination Rate | 5.3% | — |
| ARC (AI2) Challenge | — | 84.9% |
| MMLU | — | 68.8% |
| OpenBookQA | — | 88% |
| TriviaQA | — | 64% |
Multilingual DeepSeek-V3.2-Exp leads
DeepSeek-V3.2-Exp: 52.2 (#90), Phi 3 Mini 4k Instruct: 26.3 (#280)
| Benchmark | DeepSeek-V3.2-Exp | Phi 3 Mini 4k Instruct |
|---|---|---|
| LMArena Non-English | 1409 | 1021 |
| LMArena Chinese | 1461 | 1021 |
| LMArena French | 1433 | 1076 |
| LMArena German | 1440 | 1044 |
| LMArena Japanese | 1374 | 935 |
| LMArena Korean | 1371 | 905 |
| LMArena Russian | 1424 | 1022 |
| LMArena Spanish | 1440 | 1085 |
Instruction Following DeepSeek-V3.2-Exp leads
DeepSeek-V3.2-Exp: 74.5 (#93), Phi 3 Mini 4k Instruct: 47.7 (#303)
| Benchmark | DeepSeek-V3.2-Exp | Phi 3 Mini 4k Instruct |
|---|---|---|
| LMArena Instruction Following | 1413 | 1053 |
| LiveBench Instruction Following | — | 39.1% |
Long Context DeepSeek-V3.2-Exp leads
DeepSeek-V3.2-Exp: 47.6 (#16), Phi 3 Mini 4k Instruct: 31.7 (#276)
| Benchmark | DeepSeek-V3.2-Exp | Phi 3 Mini 4k Instruct |
|---|---|---|
| LMArena Longer Query | 1428 | 1044 |
| Fiction.LiveBench | 83.3% | — |
| CL-bench | 13.2% | — |
| CL-bench Life | 9.5% | — |
Writing & Preference DeepSeek-V3.2-Exp leads
DeepSeek-V3.2-Exp: 62.4 (#77), Phi 3 Mini 4k Instruct: 27.6 (#300)
| Benchmark | DeepSeek-V3.2-Exp | Phi 3 Mini 4k Instruct |
|---|---|---|
| LMArena Text | 1425 | 1073 |
| LMArena Creative Writing | 1403 | 1037 |
| LMArena Multi-Turn | 1427 | 1018 |
| EQ-Bench Creative Writing | 1515 | — |
| LiveBench Language | — | 9.2% |
Frequently asked questions
Is DeepSeek-V3.2-Exp better than Phi 3 Mini 4k Instruct?
DeepSeek-V3.2-Exp is the stronger model overall, scoring 44.3 to 27.9 on the Noometry Index.
Is DeepSeek-V3.2-Exp or Phi 3 Mini 4k Instruct better for coding?
DeepSeek-V3.2-Exp scores higher on coding benchmarks: 46.5 versus 26.6 in the Noometry coding category.
How many benchmarks do DeepSeek-V3.2-Exp and Phi 3 Mini 4k Instruct share?
18 benchmarks have published results for both models. DeepSeek-V3.2-Exp has 49 scored results on Noometry and Phi 3 Mini 4k Instruct has 35.