DeepSeek V4 Flash vs Step 5 Preview
DeepSeek V4 Flash 
Step 5 Preview Verdict
- Step 5 Preview leads on quality: 65.0 vs 62.1.
- DeepSeek V4 Flash (max) is stronger in agents & tools, coding, instruction following, maths.
- Step 5 Preview is stronger in composite, knowledge, long context, reasoning, multimodal.
- DeepSeek V4 Flash (max) is 5.4× cheaper ($0.262 vs $1.43 per 1M blended).
| Metric | DeepSeek V4 Flash (max) | Step 5 Preview |
|---|---|---|
| BenchLeader Index | 62.1 | 65.0 |
| Agents & tools score | 67.0 | – |
| Coding score | 49.1 | – |
| Composite score | 72.8 | 84.4 |
| Instruction following score | 76.7 | – |
| Knowledge score | 57.4 | 72.0 |
| Long context score | 66.3 | 70.8 |
| Maths score | 57.3 | – |
| Reasoning score | 73.9 | 82.1 |
| Multimodal score | – | 60.8 |
| Blended price $/M | $0.262 | $1.43 |
| Output speed | 212 tok/s | – |
| Time to first answer | 10.6 s | – |
| Context window | 1M | 1M |
| SciCode | 44.9% | – |
| WeirdML | 45.6% | – |
| AA Intelligence Index | 34.5 | 43.6 |
| IFBench | 79.2% | – |
| AA-LCR | 79.7% | 88.3% |
| MMMU-Pro | – | 76.4% |
| AA-Omniscience | -14.3 | 16.4 |
| Terminal-Bench Hard | 35.6% | – |
| GPQA Diamond (AA) | 90.8% | – |
| Humanity's Last Exam (AA) | 38.5% | 46.5% |
| SciCode (AA) | 50.4% | 58.9% |
| τ²-Bench Telecom (AA) | 95.0% | – |
| AIME 2026 | 95.8% | – |
| HMMT February 2026 | 93.9% | – |
| MathArena Apex | 27.1% | – |
| CritPt | 16.6% | 20.9% |
| GDPval (AA) | 47.1% | 53.6% |
| τ²-Bench Banking (AA) | 39.4% | – |
| ITBench SRE (AA) | 31.5% | – |
| Analyst Agent (AA) | 25.0% | – |
| LMCA | 35.9% | – |
| DTBench | 86.4% | – |
Data as of 2026-09-19. Best configuration of each model; every score links to its source on the model pages.
DeepSeek V4 Flash vs Step 5 Preview: questions
- Is DeepSeek V4 Flash better than Step 5 Preview?
- Step 5 Preview leads on quality: 65.0 vs 62.1. The BenchLeader Index combines every independent quality benchmark; Step 5 Preview is ahead overall as of 2026-09-19, but check the category scores for your use.
- Which is cheaper, DeepSeek V4 Flash or Step 5 Preview?
- DeepSeek V4 Flash is cheaper: $0.262 against $1.43 per million tokens, blended at three input tokens per output token.
- Which has the larger context window?
- Both accept 1M tokens of context.