DeepSeek V4 Flash vs DeepSeek V4 Pro
DeepSeek V4 Flash Verdict
- DeepSeek V4 Pro (max) leads on quality: 64.2 vs 62.1.
- DeepSeek V4 Flash (max) is stronger in instruction following, maths, reasoning.
- DeepSeek V4 Pro (max) is stronger in agents & tools, coding, composite, knowledge, long context.
- DeepSeek V4 Flash (max) is 2.1× cheaper ($0.262 vs $0.544 per 1M blended).
- DeepSeek V4 Flash (max) streams 2.6× faster (212 vs 81 tokens per second).
| Metric | DeepSeek V4 Flash (max) | DeepSeek V4 Pro (max) |
|---|---|---|
| BenchLeader Index | 62.1 | 64.2 |
| Agents & tools score | 67.0 | 70.2 |
| Coding score | 49.1 | 58.8 |
| Composite score | 72.8 | 75.0 |
| Instruction following score | 76.7 | 74.3 |
| Knowledge score | 57.4 | 58.1 |
| Long context score | 66.3 | 66.6 |
| Maths score | 57.3 | 53.7 |
| Reasoning score | 73.9 | 70.9 |
| Blended price $/M | $0.262 | $0.544 |
| Output speed | 212 tok/s | 81 tok/s |
| Time to first answer | 10.6 s | 55.6 s |
| Context window | 1M | 1M |
| GPQA Diamond | – | 89.7% |
| FrontierMath Tiers 1–3 | – | 45.3% |
| FrontierMath Tier 4 | – | 2.4% |
| OTIS Mock AIME | – | 96.7% |
| SWE-bench Verified (Epoch) | – | 77.6% |
| SimpleQA Verified | – | 47.0% |
| SciCode | 44.9% | 50.0% |
| WeirdML | 45.6% | 48.9% |
| ProofBench | – | 16.0% |
| AA Intelligence Index | 34.5 | 36.3 |
| IFBench | 79.2% | 76.5% |
| AA-LCR | 79.7% | 80.3% |
| AA-Omniscience | -14.3 | 0.8 |
| Terminal-Bench Hard | 35.6% | 46.2% |
| GPQA Diamond (AA) | 90.8% | 92.8% |
| Humanity's Last Exam (AA) | 38.5% | 41.0% |
| SciCode (AA) | 50.4% | 51.0% |
| τ²-Bench Telecom (AA) | 95.0% | 96.2% |
| LiveCodeBench | – | 87.5% |
| MMLU-Pro | – | 87.3% |
| LegalBench | – | 80.3% |
| CorpFin | – | 61.4% |
| TaxEval | – | 72.1% |
| SWE-bench (Vals) | – | 77.4% |
| GPQA Diamond (Vals) | – | 89.4% |
| Vals Index | – | 42.9 |
| AIME 2026 | 95.8% | 96.7% |
| HMMT February 2026 | 93.9% | 93.9% |
| MathArena Apex | 27.1% | 28.1% |
| CritPt | 16.6% | 18.0% |
| GDPval (AA) | 47.1% | 49.7% |
| τ²-Bench Banking (AA) | 39.4% | 39.6% |
| ITBench SRE (AA) | 31.5% | 38.3% |
| Analyst Agent (AA) | 25.0% | 18.8% |
| APEX-Agents (AA) | – | 24.3% |
| CaseLaw v2 | – | 59.4% |
| Code Migration | – | 26.2% |
| Excel Modeling Benchmark | – | 51.6% |
| Finance Agent v2 | – | 44.1% |
| Harvey's Legal Agent Benchmark | – | 3.8% |
| Legal Research Bench | – | 23.1% |
| MedCode | – | 40.5% |
| MedScribe | – | 75.1% |
| ProgramBench | – | 0.0% |
| Public Benefits Bench | – | 62.9% |
| SkillsBench | – | 51.3% |
| Vibe Code Bench v1.1 | – | 49.9% |
| Chess Puzzles | – | 20.0% |
| LMCA | 35.9% | 41.2% |
| DTBench | 86.4% | 90.7% |
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 DeepSeek V4 Pro: questions
- Is DeepSeek V4 Flash better than DeepSeek V4 Pro?
- DeepSeek V4 Pro (max) leads on quality: 64.2 vs 62.1. The BenchLeader Index combines every independent quality benchmark; DeepSeek V4 Pro (max) is ahead overall as of 2026-09-19, but check the category scores for your use.
- Is DeepSeek V4 Flash better than DeepSeek V4 Pro for coding?
- DeepSeek V4 Pro scores higher in coding (59 vs 49 on the category index, where 50 is average).
- Is DeepSeek V4 Flash better than DeepSeek V4 Pro for agentic tasks?
- DeepSeek V4 Pro scores higher in agentic tasks (70 vs 67 on the category index, where 50 is average).
- Which is cheaper, DeepSeek V4 Flash or DeepSeek V4 Pro?
- DeepSeek V4 Flash is cheaper: $0.262 against $0.544 per million tokens, blended at three input tokens per output token.
- Which is faster, DeepSeek V4 Flash or DeepSeek V4 Pro?
- DeepSeek V4 Flash streams faster: 212 against 81 output tokens per second.
- Which has the larger context window?
- Both accept 1M tokens of context.