# DeepSeek-V3.2-Speciale vs DeepSeek V4 Flash

> DeepSeek V4 Flash is the stronger model overall, scoring 53.6 to 39.1 on the Noometry Index.

- Canonical page: https://noometry.com/compare/deepseek-v3-2-speciale-vs-deepseek-v4-flash
- Last updated: 2026-10-10
- Shared benchmarks: 3

## Summary

- They share 3 benchmarks with published results for both. DeepSeek-V3.2-Speciale scores higher in 0 categories and DeepSeek V4 Flash in 3 categories; 3 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where DeepSeek V4 Flash leads 53.7 to 32.1.
- The biggest single-benchmark swing is WeirdML: 46.7% for DeepSeek-V3.2-Speciale and 63% for DeepSeek V4 Flash.
- DeepSeek V4 Flash is cheaper at $0.15 / $0.60 per million input/output tokens, against $0.58 / $1.68 for DeepSeek-V3.2-Speciale.
- DeepSeek V4 Flash accepts more context: 1M tokens versus 128K.

## Snapshot

| | DeepSeek-V3.2-Speciale | DeepSeek V4 Flash |
|---|---|---|
| Provider | DeepSeek | DeepSeek |
| Noometry Index | 39.1 | 53.6 |
| Rank | 159 | 32 |
| Context | 128K | 1M |
| Input $/M | $0.58 | $0.15 |
| Output $/M | $1.68 | $0.60 |
| Weights | Open | Open |

## Coding

- DeepSeek-V3.2-Speciale: 39.9 (#136)
- DeepSeek V4 Flash: 47.6 (#56)

| Benchmark | DeepSeek-V3.2-Speciale | DeepSeek V4 Flash |
|---|---|---|
| WeirdML | 46.7% | 63% |
| FrontierCode | — | 18.8% |
| LMArena WebDev | — | 1581 |
| SciCode | — | 49.9% |
| LMArena Coding | — | 1457 |
| ALE-Bench | — | 1,306 |

## Reasoning

- DeepSeek-V3.2-Speciale: 32.1 (#65)
- DeepSeek V4 Flash: 53.7 (#25)

| Benchmark | DeepSeek-V3.2-Speciale | DeepSeek V4 Flash |
|---|---|---|
| SimpleBench | 52.6% | 61.1% |
| ARC-AGI-2 | — | 61.4% |
| Kagi LLM Benchmark | — | 52.2% |
| NYT Connections (extended) | — | 89.6% |
| ARC-AGI-1 | — | 89% |
| CritPt | — | 16.6% |
| Chess Puzzles | — | 33% |
| LMArena Hard Prompts | — | 1443 |
| Mystery Game Puzzles | — | 34% |
| DTBench | — | 90.9% |
| LMCA | — | 41.7% |
| Epoch Capabilities Index | — | 154.51 |

## Math

- DeepSeek-V3.2-Speciale: —
- DeepSeek V4 Flash: 60.0 (#31)

| Benchmark | DeepSeek-V3.2-Speciale | DeepSeek V4 Flash |
|---|---|---|
| FrontierMath (Tiers 1-3) | — | 57.5% |
| FrontierMath Tier 4 | — | 24.4% |
| MathArena Final-Answer Competitions | — | 76.5% |
| OTIS Mock AIME 2024-2025 | — | 94.4% |
| ProofBench | — | 56% |
| LMArena Math | — | 1427 |

## Knowledge

- DeepSeek-V3.2-Speciale: —
- DeepSeek V4 Flash: 55.5 (#43)

| Benchmark | DeepSeek-V3.2-Speciale | DeepSeek V4 Flash |
|---|---|---|
| GPQA Diamond | — | 91% |
| SimpleQA Verified | — | 33.6% |
| LMArena Expert | — | 1441 |

## Multilingual

- DeepSeek-V3.2-Speciale: —
- DeepSeek V4 Flash: 53.2 (#71)

| Benchmark | DeepSeek-V3.2-Speciale | DeepSeek V4 Flash |
|---|---|---|
| LMArena Non-English | — | 1420 |
| LMArena Chinese | — | 1467 |
| LMArena French | — | 1438 |
| LMArena German | — | 1418 |
| LMArena Japanese | — | 1403 |
| LMArena Korean | — | 1383 |
| LMArena Russian | — | 1428 |
| LMArena Spanish | — | 1437 |

## Instruction Following

- DeepSeek-V3.2-Speciale: —
- DeepSeek V4 Flash: 75.1 (#76)

| Benchmark | DeepSeek-V3.2-Speciale | DeepSeek V4 Flash |
|---|---|---|
| LMArena Instruction Following | — | 1421 |

## Long Context

- DeepSeek-V3.2-Speciale: —
- DeepSeek V4 Flash: 43.8 (#81)

| Benchmark | DeepSeek-V3.2-Speciale | DeepSeek V4 Flash |
|---|---|---|
| LMArena Longer Query | — | 1435 |

## Writing & Preference

- DeepSeek-V3.2-Speciale: 46.1 (#217)
- DeepSeek V4 Flash: 64.0 (#59)

| Benchmark | DeepSeek-V3.2-Speciale | DeepSeek V4 Flash |
|---|---|---|
| EQ-Bench Creative Writing | 1276 | 1559 |
| LMArena Text | — | 1432 |
| LMArena Creative Writing | — | 1403 |
| LMArena Multi-Turn | — | 1449 |

## FAQ

### Is DeepSeek-V3.2-Speciale better than DeepSeek V4 Flash?

DeepSeek V4 Flash is the stronger model overall, scoring 53.6 to 39.1 on the Noometry Index.

### Which is cheaper, DeepSeek-V3.2-Speciale or DeepSeek V4 Flash?

DeepSeek V4 Flash is cheaper. It lists at $0.15 per million input tokens and $0.60 per million output tokens; DeepSeek-V3.2-Speciale lists at $0.58 and $1.68.

### Is DeepSeek-V3.2-Speciale or DeepSeek V4 Flash better for coding?

DeepSeek V4 Flash scores higher on coding benchmarks: 47.6 versus 39.9 in the Noometry coding category.

### Which has the bigger context window?

DeepSeek V4 Flash does, with 1M tokens against 128K.

### How many benchmarks do DeepSeek-V3.2-Speciale and DeepSeek V4 Flash share?

3 benchmarks have published results for both models. DeepSeek-V3.2-Speciale has 3 scored results on Noometry and DeepSeek V4 Flash has 41.