Model comparison
DeepSeek-V3.2-Exp vs phi-3-medium 14B
DeepSeek-V3.2-Exp is the stronger model overall, scoring 44.3 to 29.7 on the Noometry Index.
Last verified . 2 shared benchmarks.
DeepSeek
MicrosoftSummary
- They share 2 benchmarks with published results for both. DeepSeek-V3.2-Exp scores higher in 3 categories and phi-3-medium 14B in 0 categories; 3 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where DeepSeek-V3.2-Exp leads 51.7 to 9.1.
- The biggest single-benchmark swing is GPQA Diamond: 83.4% for DeepSeek-V3.2-Exp and 27.6% for phi-3-medium 14B.
Side by side
| DeepSeek-V3.2-Exp | phi-3-medium 14B | |
|---|---|---|
| Provider | DeepSeek | Microsoft |
| Noometry Index | 44.3 | 29.7 |
| 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 | 13 |
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-medium 14B: 36.8 (#201)
| Benchmark | DeepSeek-V3.2-Exp | phi-3-medium 14B |
|---|---|---|
| SWE-bench Verified (bash only) | 70% | — |
| Aider Polyglot | 74.2% | — |
| LMArena WebDev | 1362 | — |
| SWE-bench Multilingual | 59% | — |
| SciCode | 38.9% | — |
| WeirdML | 39.5% | — |
| BigCodeBench Instruct | — | 37.6% |
| LMArena Coding | 1454 | — |
| BigCodeBench Complete | — | 48.7% |
Agentic & Tool Use Not comparable
DeepSeek-V3.2-Exp: 32.7 (#59), phi-3-medium 14B: —
| Benchmark | DeepSeek-V3.2-Exp | phi-3-medium 14B |
|---|---|---|
| Terminal-Bench | 39.6% | — |
| APEX-Agents | 21.3% | — |
| Berkeley Function Calling Leaderboard | 56.7% | — |
| TheAgentCompany | 42.9% | — |
| Vending-Bench 2 | 1,034 | — |
Reasoning Not comparable
DeepSeek-V3.2-Exp: 22.1 (#208), phi-3-medium 14B: —
| Benchmark | DeepSeek-V3.2-Exp | phi-3-medium 14B |
|---|---|---|
| Epoch Capabilities Index | 146.27 | 121.23 |
| ARC-AGI-2 | 4% | — |
| Kagi LLM Benchmark | 52.2% | — |
| NYT Connections (extended) | 36.7% | — |
| ARC-AGI-1 | 57% | — |
| CritPt | 2.9% | — |
| Chess Puzzles | 14% | — |
| Thematic Generalization | 65% | — |
| LMArena Hard Prompts | 1434 | — |
| DTBench | 87.7% | — |
| LMCA | 29.1% | — |
| Adversarial NLI | — | 55.8% |
| BIG-Bench Hard | — | 81.4% |
| HellaSwag | — | 82.4% |
| WinoGrande | — | 81.5% |
Math DeepSeek-V3.2-Exp leads
DeepSeek-V3.2-Exp: 41.7 (#87), phi-3-medium 14B: 27.3 (#250)
| Benchmark | DeepSeek-V3.2-Exp | phi-3-medium 14B |
|---|---|---|
| MathArena Final-Answer Competitions | 57.7% | — |
| OTIS Mock AIME 2024-2025 | 87.8% | — |
| ProofBench | 8% | — |
| LMArena Math | 1435 | — |
| MATH Level 5 | — | 17.6% |
| 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-medium 14B: 9.1 (#306)
| Benchmark | DeepSeek-V3.2-Exp | phi-3-medium 14B |
|---|---|---|
| GPQA Diamond | 83.4% | 27.6% |
| Vectara Hallucination Rate | 5.3% | — |
| LMArena Expert | 1436 | — |
| ARC (AI2) Challenge | — | 91.6% |
| MMLU | — | 78% |
| OpenBookQA | — | 87.4% |
| TriviaQA | — | 73.9% |
Multilingual Not comparable
DeepSeek-V3.2-Exp: 52.2 (#90), phi-3-medium 14B: —
| Benchmark | DeepSeek-V3.2-Exp | phi-3-medium 14B |
|---|---|---|
| LMArena Non-English | 1409 | — |
| LMArena Chinese | 1461 | — |
| LMArena French | 1433 | — |
| LMArena German | 1440 | — |
| LMArena Japanese | 1374 | — |
| LMArena Korean | 1371 | — |
| LMArena Russian | 1424 | — |
| LMArena Spanish | 1440 | — |
Instruction Following Not comparable
DeepSeek-V3.2-Exp: 74.5 (#93), phi-3-medium 14B: —
| Benchmark | DeepSeek-V3.2-Exp | phi-3-medium 14B |
|---|---|---|
| LMArena Instruction Following | 1413 | — |
Long Context Not comparable
DeepSeek-V3.2-Exp: 47.6 (#16), phi-3-medium 14B: —
| Benchmark | DeepSeek-V3.2-Exp | phi-3-medium 14B |
|---|---|---|
| Fiction.LiveBench | 83.3% | — |
| CL-bench | 13.2% | — |
| CL-bench Life | 9.5% | — |
| LMArena Longer Query | 1428 | — |
Writing & Preference Not comparable
DeepSeek-V3.2-Exp: 62.4 (#77), phi-3-medium 14B: —
| Benchmark | DeepSeek-V3.2-Exp | phi-3-medium 14B |
|---|---|---|
| LMArena Text | 1425 | — |
| LMArena Creative Writing | 1403 | — |
| EQ-Bench Creative Writing | 1515 | — |
| LMArena Multi-Turn | 1427 | — |
Frequently asked questions
Is DeepSeek-V3.2-Exp better than phi-3-medium 14B?
DeepSeek-V3.2-Exp is the stronger model overall, scoring 44.3 to 29.7 on the Noometry Index.
Is DeepSeek-V3.2-Exp or phi-3-medium 14B better for coding?
DeepSeek-V3.2-Exp scores higher on coding benchmarks: 46.5 versus 36.8 in the Noometry coding category.
How many benchmarks do DeepSeek-V3.2-Exp and phi-3-medium 14B share?
2 benchmarks have published results for both models. DeepSeek-V3.2-Exp has 49 scored results on Noometry and phi-3-medium 14B has 13.