# DeepSeek V4 Flash vs gpt-oss-20b

> DeepSeek V4 Flash is the stronger model overall, scoring 53.6 to 32.5 on the Noometry Index. gpt-oss-20b costs 7.3× less per token, which makes it the better buy when DeepSeek V4 Flash's lead doesn't matter for your workload.

- Canonical page: https://noometry.com/compare/deepseek-v4-flash-vs-gpt-oss-20b
- Last updated: 2026-10-10
- Shared benchmarks: 28

## Summary

- They share 28 benchmarks with published results for both. DeepSeek V4 Flash scores higher in 8 categories and gpt-oss-20b in 0 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where DeepSeek V4 Flash leads 53.7 to 19.3.
- The biggest single-benchmark swing is GPQA Diamond: 91% for DeepSeek V4 Flash and 60.8% for gpt-oss-20b.
- gpt-oss-20b is cheaper at $0.018 / $0.09 per million input/output tokens, against $0.15 / $0.60 for DeepSeek V4 Flash.
- DeepSeek V4 Flash accepts more context: 1M tokens versus 131K.

## Snapshot

| | DeepSeek V4 Flash | gpt-oss-20b |
|---|---|---|
| Provider | DeepSeek | OpenAI |
| Noometry Index | 53.6 | 32.5 |
| Rank | 35 | 255 |
| Context | 1M | 131K |
| Input $/M | $0.15 | $0.018 |
| Output $/M | $0.60 | $0.09 |
| Weights | Open | Open |

## Coding

- DeepSeek V4 Flash: 47.9 (#59)
- gpt-oss-20b: 37.6 (#192)

| Benchmark | DeepSeek V4 Flash | gpt-oss-20b |
|---|---|---|
| SciCode | 49.9% | 34.4% |
| WeirdML | 63% | 40.9% |
| LMArena Coding | 1457 | 1306 |
| ALE-Bench | 1,306 | 566.05 |
| FrontierCode | 18.8% | — |
| LMArena WebDev | 1582 | — |

## Agentic & Tool Use

- DeepSeek V4 Flash: —
- gpt-oss-20b: 9.3 (#154)

| Benchmark | DeepSeek V4 Flash | gpt-oss-20b |
|---|---|---|
| Terminal-Bench | — | 3.4% |

## Reasoning

- DeepSeek V4 Flash: 53.7 (#30)
- gpt-oss-20b: 19.3 (#261)

| Benchmark | DeepSeek V4 Flash | gpt-oss-20b |
|---|---|---|
| Kagi LLM Benchmark | 52.2% | 53.2% |
| CritPt | 16.6% | 1.4% |
| Chess Puzzles | 33% | 4% |
| LMArena Hard Prompts | 1444 | 1274 |
| DTBench | 90.9% | 68% |
| LMCA | 41.7% | 14.5% |
| Epoch Capabilities Index | 154.49 | 137.82 |
| ARC-AGI-2 | 61.4% | — |
| SimpleBench | 61.1% | — |
| NYT Connections (extended) | 89.6% | — |
| ARC-AGI-1 | 89% | — |
| Mystery Game Puzzles | 34% | — |

## Math

- DeepSeek V4 Flash: 60.3 (#37)
- gpt-oss-20b: 39.4 (#103)

| Benchmark | DeepSeek V4 Flash | gpt-oss-20b |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 94.4% | 65.3% |
| LMArena Math | 1427 | 1317 |
| FrontierMath (Tiers 1-3) | 57.5% | — |
| FrontierMath Tier 4 | 24.4% | — |
| MathArena Final-Answer Competitions | 76.5% | — |
| ProofBench | 56% | — |
| Omni-MATH | — | 56.5% |

## Knowledge

- DeepSeek V4 Flash: 55.4 (#48)
- gpt-oss-20b: 34.6 (#195)

| Benchmark | DeepSeek V4 Flash | gpt-oss-20b |
|---|---|---|
| GPQA Diamond | 91% | 60.8% |
| LMArena Expert | 1441 | 1258 |
| SimpleQA Verified | 33.6% | — |
| MMLU-Pro | — | 74% |
| GPQA (HELM) | — | 59.4% |

## Multilingual

- DeepSeek V4 Flash: 53.0 (#72)
- gpt-oss-20b: 42.2 (#197)

| Benchmark | DeepSeek V4 Flash | gpt-oss-20b |
|---|---|---|
| LMArena Non-English | 1420 | 1268 |
| LMArena Chinese | 1468 | 1314 |
| LMArena German | 1418 | 1255 |
| LMArena Japanese | 1406 | 1244 |
| LMArena Korean | 1384 | 1236 |
| LMArena Russian | 1428 | 1278 |
| LMArena Spanish | 1436 | 1267 |
| LMArena French | 1439 | — |

## Instruction Following

- DeepSeek V4 Flash: 74.9 (#81)
- gpt-oss-20b: 61.8 (#240)

| Benchmark | DeepSeek V4 Flash | gpt-oss-20b |
|---|---|---|
| LMArena Instruction Following | 1421 | 1236 |
| IFEval | — | 73.2% |

## Long Context

- DeepSeek V4 Flash: 43.8 (#85)
- gpt-oss-20b: 37.9 (#209)

| Benchmark | DeepSeek V4 Flash | gpt-oss-20b |
|---|---|---|
| LMArena Longer Query | 1434 | 1250 |

## Writing & Preference

- DeepSeek V4 Flash: 63.8 (#61)
- gpt-oss-20b: 35.5 (#265)

| Benchmark | DeepSeek V4 Flash | gpt-oss-20b |
|---|---|---|
| LMArena Text | 1432 | 1287 |
| LMArena Creative Writing | 1403 | 1201 |
| EQ-Bench Creative Writing | 1559 | 666 |
| LMArena Multi-Turn | 1449 | 1268 |
| WildBench | — | 73.7% |

## FAQ

### Is DeepSeek V4 Flash better than gpt-oss-20b?

DeepSeek V4 Flash is the stronger model overall, scoring 53.6 to 32.5 on the Noometry Index. gpt-oss-20b costs 7.3× less per token, which makes it the better buy when DeepSeek V4 Flash's lead doesn't matter for your workload.

### Which is cheaper, DeepSeek V4 Flash or gpt-oss-20b?

gpt-oss-20b is cheaper. It lists at $0.018 per million input tokens and $0.09 per million output tokens; DeepSeek V4 Flash lists at $0.15 and $0.60.

### Is DeepSeek V4 Flash or gpt-oss-20b better for coding?

DeepSeek V4 Flash scores higher on coding benchmarks: 47.9 versus 37.6 in the Noometry coding category.

### Which has the bigger context window?

DeepSeek V4 Flash does, with 1M tokens against 131K.

### How many benchmarks do DeepSeek V4 Flash and gpt-oss-20b share?

28 benchmarks have published results for both models. DeepSeek V4 Flash has 41 scored results on Noometry and gpt-oss-20b has 34.