# DeepSeek V4 Flash vs Inkling-Small

> DeepSeek V4 Flash is the stronger model overall, scoring 53.6 to 46.5 on the Noometry Index.

- Canonical page: https://noometry.com/compare/deepseek-v4-flash-vs-inkling-small
- Last updated: 2026-10-11
- Shared benchmarks: 32

## Summary

- They share 32 benchmarks with published results for both. DeepSeek V4 Flash scores higher in 8 categories and Inkling-Small in 0 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in math, where DeepSeek V4 Flash leads 60.3 to 45.1.
- The biggest single-benchmark swing is ProofBench: 56% for DeepSeek V4 Flash and 6% for Inkling-Small.
- DeepSeek V4 Flash is cheaper at $0.15 / $0.60 per million input/output tokens, against $0.45 / $1.20 for Inkling-Small.
- DeepSeek V4 Flash accepts more context: 1M tokens versus 524K.

## Snapshot

| | DeepSeek V4 Flash | Inkling-Small |
|---|---|---|
| Provider | DeepSeek | Thinking Machines Lab |
| Noometry Index | 53.6 | 46.5 |
| Rank | 35 | 63 |
| Context | 1M | 524K |
| Input $/M | $0.15 | $0.45 |
| Output $/M | $0.60 | $1.20 |
| Weights | Open | Open |

## Coding

- DeepSeek V4 Flash: 47.9 (#59)
- Inkling-Small: 43.6 (#85)

| Benchmark | DeepSeek V4 Flash | Inkling-Small |
|---|---|---|
| LMArena WebDev | 1582 | 1409 |
| SciCode | 49.9% | 48.7% |
| LMArena Coding | 1457 | 1451 |
| FrontierCode | 18.8% | — |
| WeirdML | 63% | — |
| ALE-Bench | 1,306 | — |

## Reasoning

- DeepSeek V4 Flash: 53.7 (#30)
- Inkling-Small: 38.6 (#63)

| Benchmark | DeepSeek V4 Flash | Inkling-Small |
|---|---|---|
| ARC-AGI-2 | 61.4% | 40.1% |
| ARC-AGI-1 | 89% | 84% |
| CritPt | 16.6% | 8.3% |
| Chess Puzzles | 33% | 18% |
| LMArena Hard Prompts | 1444 | 1423 |
| Mystery Game Puzzles | 34% | 6% |
| Epoch Capabilities Index | 154.49 | 150.15 |
| SimpleBench | 61.1% | — |
| Kagi LLM Benchmark | 52.2% | — |
| NYT Connections (extended) | 89.6% | — |
| DTBench | 90.9% | — |
| LMCA | 41.7% | — |

## Math

- DeepSeek V4 Flash: 60.3 (#37)
- Inkling-Small: 45.1 (#77)

| Benchmark | DeepSeek V4 Flash | Inkling-Small |
|---|---|---|
| FrontierMath (Tiers 1-3) | 57.5% | 46.3% |
| FrontierMath Tier 4 | 24.4% | 17.1% |
| OTIS Mock AIME 2024-2025 | 94.4% | 90% |
| ProofBench | 56% | 6% |
| LMArena Math | 1427 | 1459 |
| MathArena Final-Answer Competitions | 76.5% | — |

## Knowledge

- DeepSeek V4 Flash: 55.4 (#48)
- Inkling-Small: 48.2 (#77)

| Benchmark | DeepSeek V4 Flash | Inkling-Small |
|---|---|---|
| GPQA Diamond | 91% | 88.5% |
| SimpleQA Verified | 33.6% | 19.1% |
| LMArena Expert | 1441 | 1442 |

## Multimodal

- DeepSeek V4 Flash: —
- Inkling-Small: 39.1 (#62)

| Benchmark | DeepSeek V4 Flash | Inkling-Small |
|---|---|---|
| LMArena Vision | — | 1235 |

## Multilingual

- DeepSeek V4 Flash: 53.0 (#72)
- Inkling-Small: 51.7 (#104)

| Benchmark | DeepSeek V4 Flash | Inkling-Small |
|---|---|---|
| LMArena Non-English | 1420 | 1402 |
| LMArena Chinese | 1468 | 1465 |
| LMArena French | 1439 | 1436 |
| LMArena German | 1418 | 1405 |
| LMArena Japanese | 1406 | 1405 |
| LMArena Korean | 1384 | 1363 |
| LMArena Russian | 1428 | 1391 |
| LMArena Spanish | 1436 | 1428 |

## Instruction Following

- DeepSeek V4 Flash: 74.9 (#81)
- Inkling-Small: 73.8 (#114)

| Benchmark | DeepSeek V4 Flash | Inkling-Small |
|---|---|---|
| LMArena Instruction Following | 1421 | 1399 |

## Long Context

- DeepSeek V4 Flash: 43.8 (#85)
- Inkling-Small: 42.7 (#118)

| Benchmark | DeepSeek V4 Flash | Inkling-Small |
|---|---|---|
| LMArena Longer Query | 1434 | 1401 |

## Writing & Preference

- DeepSeek V4 Flash: 63.8 (#61)
- Inkling-Small: 59.6 (#107)

| Benchmark | DeepSeek V4 Flash | Inkling-Small |
|---|---|---|
| LMArena Text | 1432 | 1414 |
| LMArena Creative Writing | 1403 | 1331 |
| EQ-Bench Creative Writing | 1559 | 1491 |
| LMArena Multi-Turn | 1449 | 1418 |

## FAQ

### Is DeepSeek V4 Flash better than Inkling-Small?

DeepSeek V4 Flash is the stronger model overall, scoring 53.6 to 46.5 on the Noometry Index.

### Which is cheaper, DeepSeek V4 Flash or Inkling-Small?

DeepSeek V4 Flash is cheaper. It lists at $0.15 per million input tokens and $0.60 per million output tokens; Inkling-Small lists at $0.45 and $1.20.

### Is DeepSeek V4 Flash or Inkling-Small better for coding?

DeepSeek V4 Flash scores higher on coding benchmarks: 47.9 versus 43.6 in the Noometry coding category.

### Which has the bigger context window?

DeepSeek V4 Flash does, with 1M tokens against 524K.

### How many benchmarks do DeepSeek V4 Flash and Inkling-Small share?

32 benchmarks have published results for both models. DeepSeek V4 Flash has 41 scored results on Noometry and Inkling-Small has 33.