Model comparison
DeepSeek-V3.2-Speciale vs Gemini 3.7 Flash
Gemini 3.7 Flash is the stronger model overall, scoring 59.8 to 39.7 on the Noometry Index. DeepSeek-V3.2-Speciale costs 1.8× less per token, which makes it the better buy when Gemini 3.7 Flash's lead doesn't matter for your workload.
Last verified . 1 shared benchmarks.
DeepSeek
GoogleSummary
- They share 1 benchmark with published results for both. DeepSeek-V3.2-Speciale scores higher in 0 categories and Gemini 3.7 Flash in 3 categories; 3 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where Gemini 3.7 Flash leads 70.0 to 32.9.
- DeepSeek-V3.2-Speciale is cheaper at $0.58 / $1.68 per million input/output tokens, against $0.75 / $3.75 for Gemini 3.7 Flash.
- Gemini 3.7 Flash accepts more context: 1.05M tokens versus 128K.
- DeepSeek-V3.2-Speciale has downloadable open weights; the other is API-only.
Side by side
| DeepSeek-V3.2-Speciale | Gemini 3.7 Flash | |
|---|---|---|
| Provider | DeepSeek | |
| Noometry Index | 39.7 | 59.8 |
| Released | 2025-12-01 | 2026-08-13 |
| Weights | Open | Proprietary |
| Context window | 128K | 1.05M |
| Max output | 128K | 66K |
| Input $ / M tokens | $0.58 | $0.75 |
| Output $ / M tokens | $1.68 | $3.75 |
| Results tracked | 3 | 44 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Gemini 3.7 Flash leads
DeepSeek-V3.2-Speciale: 40.4 (#140), Gemini 3.7 Flash: 56.2 (#22)
| Benchmark | DeepSeek-V3.2-Speciale | Gemini 3.7 Flash |
|---|---|---|
| DeepSWE | — | 65.5% |
| FrontierCode | — | 43.6% |
| LMArena WebDev | — | 1592 |
| FrontierSWE | — | 20.3% |
| SciCode | — | 59.8% |
| WeirdML | 46.7% | — |
| LMArena Coding | — | 1497 |
| ALE-Bench | — | 904.3 |
Agentic & Tool Use Not comparable
DeepSeek-V3.2-Speciale: —, Gemini 3.7 Flash: 42.1 (#19)
| Benchmark | DeepSeek-V3.2-Speciale | Gemini 3.7 Flash |
|---|---|---|
| APEX-Agents | — | 67.8% |
| Remote Labor Index | — | 5% |
| GDP.pdf | — | 23.8% |
Reasoning Gemini 3.7 Flash leads
DeepSeek-V3.2-Speciale: 32.9 (#73), Gemini 3.7 Flash: 70.0 (#15)
| Benchmark | DeepSeek-V3.2-Speciale | Gemini 3.7 Flash |
|---|---|---|
| ARC-AGI-2 | — | 84.6% |
| SimpleBench | 52.6% | — |
| NYT Connections (extended) | — | 94% |
| ARC-AGI-1 | — | 95.5% |
| CritPt | — | 14.3% |
| Chess Puzzles | — | 47% |
| LMArena Hard Prompts | — | 1494 |
| Mystery Game Puzzles | — | 37% |
| DTBench | — | 96.8% |
| LMCA | — | 50.4% |
| Epoch Capabilities Index | — | 157.27 |
Math Not comparable
DeepSeek-V3.2-Speciale: —, Gemini 3.7 Flash: 69.6 (#23)
| Benchmark | DeepSeek-V3.2-Speciale | Gemini 3.7 Flash |
|---|---|---|
| FrontierMath (Tiers 1-3) | — | 71.6% |
| FrontierMath Tier 4 | — | 36.6% |
| OTIS Mock AIME 2024-2025 | — | 97.2% |
| ProofBench | — | 58% |
| LMArena Math | — | 1507 |
Knowledge Not comparable
DeepSeek-V3.2-Speciale: —, Gemini 3.7 Flash: 69.7 (#5)
| Benchmark | DeepSeek-V3.2-Speciale | Gemini 3.7 Flash |
|---|---|---|
| GPQA Diamond | — | 94.8% |
| SimpleQA Verified | — | 69.2% |
| LMArena Expert | — | 1508 |
Multimodal Not comparable
DeepSeek-V3.2-Speciale: —, Gemini 3.7 Flash: 37.3 (#73)
| Benchmark | DeepSeek-V3.2-Speciale | Gemini 3.7 Flash |
|---|---|---|
| LMArena Vision | — | 1316 |
| Furniture Assembly | — | 26.7% |
Multilingual Not comparable
DeepSeek-V3.2-Speciale: —, Gemini 3.7 Flash: 57.6 (#7)
| Benchmark | DeepSeek-V3.2-Speciale | Gemini 3.7 Flash |
|---|---|---|
| LMArena Non-English | — | 1484 |
| LMArena Chinese | — | 1548 |
| LMArena French | — | 1505 |
| LMArena German | — | 1498 |
| LMArena Japanese | — | 1512 |
| LMArena Korean | — | 1483 |
| LMArena Russian | — | 1516 |
| LMArena Spanish | — | 1503 |
Instruction Following Not comparable
DeepSeek-V3.2-Speciale: —, Gemini 3.7 Flash: 77.7 (#15)
| Benchmark | DeepSeek-V3.2-Speciale | Gemini 3.7 Flash |
|---|---|---|
| LMArena Instruction Following | — | 1483 |
Long Context Not comparable
DeepSeek-V3.2-Speciale: —, Gemini 3.7 Flash: 45.7 (#30)
| Benchmark | DeepSeek-V3.2-Speciale | Gemini 3.7 Flash |
|---|---|---|
| LMArena Longer Query | — | 1492 |
Writing & Preference Gemini 3.7 Flash leads
DeepSeek-V3.2-Speciale: 46.0 (#222), Gemini 3.7 Flash: 71.2 (#20)
| Benchmark | DeepSeek-V3.2-Speciale | Gemini 3.7 Flash |
|---|---|---|
| EQ-Bench Creative Writing | 1276 | 1723 |
| LMArena Text | — | 1486 |
| LMArena Creative Writing | — | 1490 |
| LMArena Multi-Turn | — | 1489 |
Frequently asked questions
Is DeepSeek-V3.2-Speciale better than Gemini 3.7 Flash?
Gemini 3.7 Flash is the stronger model overall, scoring 59.8 to 39.7 on the Noometry Index. DeepSeek-V3.2-Speciale costs 1.8× less per token, which makes it the better buy when Gemini 3.7 Flash's lead doesn't matter for your workload.
Which is cheaper, DeepSeek-V3.2-Speciale or Gemini 3.7 Flash?
DeepSeek-V3.2-Speciale is cheaper. It lists at $0.58 per million input tokens and $1.68 per million output tokens; Gemini 3.7 Flash lists at $0.75 and $3.75.
Is DeepSeek-V3.2-Speciale or Gemini 3.7 Flash better for coding?
Gemini 3.7 Flash scores higher on coding benchmarks: 56.2 versus 40.4 in the Noometry coding category.
Which has the bigger context window?
Gemini 3.7 Flash does, with 1.05M tokens against 128K.
How many benchmarks do DeepSeek-V3.2-Speciale and Gemini 3.7 Flash share?
1 benchmark has published results for both models. DeepSeek-V3.2-Speciale has 3 scored results on Noometry and Gemini 3.7 Flash has 44.