Model comparison
DeepSeek-V3.2-Speciale vs Phi 3 Mini 128k Instruct
DeepSeek-V3.2-Speciale is the stronger model overall, scoring 39.7 to 29.7 on the Noometry Index.
Last verified . 0 shared benchmarks.
DeepSeek
MicrosoftSummary
- The widest gap is in writing & preference, where DeepSeek-V3.2-Speciale leads 46.0 to 27.1.
Side by side
| DeepSeek-V3.2-Speciale | Phi 3 Mini 128k Instruct | |
|---|---|---|
| Provider | DeepSeek | Microsoft |
| Noometry Index | 39.7 | 29.7 |
| Released | 2025-12-01 | 2024-04-23 |
| Weights | Open | Open |
| Context window | 128K | — |
| Max output | 128K | — |
| Input $ / M tokens | $0.58 | — |
| Output $ / M tokens | $1.68 | — |
| Results tracked | 3 | 19 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding DeepSeek-V3.2-Speciale leads
DeepSeek-V3.2-Speciale: 40.4 (#140), Phi 3 Mini 128k Instruct: 28.8 (#312)
| Benchmark | DeepSeek-V3.2-Speciale | Phi 3 Mini 128k Instruct |
|---|---|---|
| WeirdML | 46.7% | — |
| BigCodeBench Instruct | — | 29.6% |
| LMArena Coding | — | 1039 |
| BigCodeBench Complete | — | 40.6% |
Reasoning DeepSeek-V3.2-Speciale leads
DeepSeek-V3.2-Speciale: 32.9 (#73), Phi 3 Mini 128k Instruct: 19.6 (#256)
| Benchmark | DeepSeek-V3.2-Speciale | Phi 3 Mini 128k Instruct |
|---|---|---|
| SimpleBench | 52.6% | — |
| LMArena Hard Prompts | — | 1028 |
Math Not comparable
DeepSeek-V3.2-Speciale: —, Phi 3 Mini 128k Instruct: 31.6 (#222)
| Benchmark | DeepSeek-V3.2-Speciale | Phi 3 Mini 128k Instruct |
|---|---|---|
| LMArena Math | — | 1089 |
Knowledge Not comparable
DeepSeek-V3.2-Speciale: —, Phi 3 Mini 128k Instruct: 26.8 (#254)
| Benchmark | DeepSeek-V3.2-Speciale | Phi 3 Mini 128k Instruct |
|---|---|---|
| LMArena Expert | — | 984 |
Multilingual Not comparable
DeepSeek-V3.2-Speciale: —, Phi 3 Mini 128k Instruct: 25.2 (#285)
| Benchmark | DeepSeek-V3.2-Speciale | Phi 3 Mini 128k Instruct |
|---|---|---|
| LMArena Non-English | — | 1000 |
| LMArena Chinese | — | 1016 |
| LMArena French | — | 1039 |
| LMArena German | — | 1006 |
| LMArena Japanese | — | 899 |
| LMArena Korean | — | 856 |
| LMArena Russian | — | 1004 |
| LMArena Spanish | — | 1059 |
Instruction Following Not comparable
DeepSeek-V3.2-Speciale: —, Phi 3 Mini 128k Instruct: 51.8 (#294)
| Benchmark | DeepSeek-V3.2-Speciale | Phi 3 Mini 128k Instruct |
|---|---|---|
| LMArena Instruction Following | — | 1023 |
Long Context Not comparable
DeepSeek-V3.2-Speciale: —, Phi 3 Mini 128k Instruct: 30.4 (#289)
| Benchmark | DeepSeek-V3.2-Speciale | Phi 3 Mini 128k Instruct |
|---|---|---|
| LMArena Longer Query | — | 996 |
Writing & Preference DeepSeek-V3.2-Speciale leads
DeepSeek-V3.2-Speciale: 46.0 (#222), Phi 3 Mini 128k Instruct: 27.1 (#301)
| Benchmark | DeepSeek-V3.2-Speciale | Phi 3 Mini 128k Instruct |
|---|---|---|
| LMArena Text | — | 1050 |
| LMArena Creative Writing | — | 1024 |
| EQ-Bench Creative Writing | 1276 | — |
| LMArena Multi-Turn | — | 989 |
Frequently asked questions
Is DeepSeek-V3.2-Speciale better than Phi 3 Mini 128k Instruct?
DeepSeek-V3.2-Speciale is the stronger model overall, scoring 39.7 to 29.7 on the Noometry Index.
Is DeepSeek-V3.2-Speciale or Phi 3 Mini 128k Instruct better for coding?
DeepSeek-V3.2-Speciale scores higher on coding benchmarks: 40.4 versus 28.8 in the Noometry coding category.
How many benchmarks do DeepSeek-V3.2-Speciale and Phi 3 Mini 128k Instruct share?
0 benchmarks have published results for both models. DeepSeek-V3.2-Speciale has 3 scored results on Noometry and Phi 3 Mini 128k Instruct has 19.