DeepSeek V4.1 Flash vs Hy4 preview
DeepSeek V4.1 Flash leads Hy4 preview on coding, 62.5 to 44.3 on the Vibe Coding Index, and DeepSeek V4.1 Flash costs 2.9x less per output token ($0.78 vs $2.2509 per million). Updated October 21, 2026; 22 benchmarks have a result for both.
vs boards won of 22 shared
DeepSeek V4.1 Flash or Hy4 preview: which to pick
Pick DeepSeek V4.1 Flash for coding: it scores 62.5 on the Vibe Coding Index against 44.3 for Hy4 preview.
Pick Hy4 preview for work measured by LMArena WebDev or Code migration (Vals AI), where it leads.
Model ADeepSeek V4.1 FlashDeepSeek62.5/1005 independent benchmarksNo cost per solved taskModel BHy4 previewTencent44.3/1003 independent benchmarksNo cost per solved task
- Shared results
- 22
- DeepSeek V4.1 Flash only
- 79
- Hy4 preview only
- 2
- Like-for-like categories
- 5 of 9
Only benchmarks that tested both models count toward a pick; a result only one model has is shown but not compared.
Which one for your work
Each card is decided by one number both models carry. Where only one has it, the card says so and makes no call.
Coding
Vibe Coding Index, six public coding benchmarks
DeepSeek V4.1 Flash
DeepSeek V4.1 Flash scores 62.5 against 44.3.
5 and 3 benchmarks behind the scores
Intelligence
AA Intelligence Index
No call
Only DeepSeek V4.1 Flash has a AA Intelligence Index result, 24.7.
Directional only
Web design
Design Arena: websites
No call
Only DeepSeek V4.1 Flash has a Design Arena: websites result, 1309.
Directional only
General
Vals Index
No call
Only Hy4 preview has a Vals Index result, 53.2%.
Directional only
Long documents
Listed context window
No call
Both take 1.05M tokens.
Listed by OpenRouter
Everyday cost
1K input + 500 output tokens
DeepSeek V4.1 Flash
DeepSeek V4.1 Flash $0.0006, Hy4 preview $0.0019 per turn.
List prices
At a glance
- Coding index
- DeepSeek V4.1 Flash 62.5 from 5 benchmarks, Hy4 preview 44.3 from 3 benchmarks.
- Widest category
- Math, 21.0 points apart on average over 1 shared benchmark.
- Widest single gap
- ProofBench: DeepSeek V4.1 Flash 54.0% against Hy4 preview 75.0%, 21.0 points apart on a 0 to 100 scale.
- Output price
- DeepSeek V4.1 Flash $0.78 against Hy4 preview $2.2509 per million output tokens, 2.9x apart.
- Context window
- Both take 1.05M tokens.
Benchmark by benchmark
First the score: boards each model leads, per category, where both have a result. Then every board, a category to a fold; bars are each result on a 0 to 100 scale, so compare the two in a row, not one row with another.
| Category | DeepSeek V4.1 Flash | Bars | Hy4 preview | Ties |
|---|---|---|---|---|
| Coding | 5 | 5 | ||
| Agentic work | 0 | 7 | ||
| Math | 0 | 1 | ||
| Knowledge | 0 | 2 | ||
| General | 1 | 1 | ||
| All shared boards | 6 | 16 |
CodingLike-for-likeSix of these boards feed the Vibe Coding Index55
| Benchmark | DeepSeek V4.1 Flash | Hy4 preview | Leader |
|---|---|---|---|
| LiveBench coding | 90.2% | Not measured | Only DeepSeek V4.1 Flash measured |
| LMArena WebDev | 1618 | 1631 | Hy4 preview |
| Terminal-Bench 4 (Vals AI) | 19.7% | 8.1% | DeepSeek V4.1 Flash |
| Vibe Code Bench (Vals AI) | 84.7% | 77.5% | DeepSeek V4.1 Flash |
| Code migration (Vals AI) | 45.6% | 47.4% | Hy4 preview |
| ProgramBench (Vals AI) | 0.5% | 0.0% | DeepSeek V4.1 Flash |
| IOI (Vals AI) | 40.3% | 59.3% | Hy4 preview |
| Terminal-Bench 2.1 (Vals AI) | 74.5% | 55.1% | DeepSeek V4.1 Flash |
| SWE-bench Multimodal (Vals AI) | 46.5% | Not measured | Only DeepSeek V4.1 Flash measured |
| Vibe Code Bench 1-100 (Vals AI) | 16.4% | Not measured | Only DeepSeek V4.1 Flash measured |
| SkillsBench (Vals AI) | 69.8% | Not measured | Only DeepSeek V4.1 Flash measured |
| CyberBench (Vals AI) | 73.7% | 65.4% | DeepSeek V4.1 Flash |
| SciCode | 51.9% | Not measured | Only DeepSeek V4.1 Flash measured |
| FrontierCode | 43.9% | Not measured | Only DeepSeek V4.1 Flash measured |
| ALE-Bench | 1092.3 | Not measured | Only DeepSeek V4.1 Flash measured |
| LMArena Text: coding | 1502 | Not measured | Only DeepSeek V4.1 Flash measured |
| LMArena Text: software and IT | 1496 | Not measured | Only DeepSeek V4.1 Flash measured |
| LMArena WebDev: React | 1628 | 1653 | Hy4 preview |
| LMArena WebDev: HTML | 1612 | 1620 | Hy4 preview |
| LMArena WebDev: from an image | 1606 | Not measured | Only DeepSeek V4.1 Flash measured |
| Neither model has a result on the other 19 boards here | |||
IntelligenceDirectional onlyPublic boards, shown as published, kept out of the index2 boards, none shared
| Benchmark | DeepSeek V4.1 Flash | Hy4 preview | Leader |
|---|---|---|---|
| AA Intelligence Index | 24.7 | Not measured | Only DeepSeek V4.1 Flash measured |
| Epoch Capabilities Index | 155.9 | Not measured | Only DeepSeek V4.1 Flash measured |
Agentic workLike-for-likePublic boards, shown as published, kept out of the index07
| Benchmark | DeepSeek V4.1 Flash | Hy4 preview | Leader |
|---|---|---|---|
| LiveBench research | 86.8% | Not measured | Only DeepSeek V4.1 Flash measured |
| LiveBench browser | 84.0% | Not measured | Only DeepSeek V4.1 Flash measured |
| LiveBench long agentic | 40.1% | Not measured | Only DeepSeek V4.1 Flash measured |
| SREBench (Vals AI) | 0.8% | 2.3% | Hy4 preview |
| Finance Agent (Vals AI) | 53.5% | 55.1% | Hy4 preview |
| Tax Agent Bench (Vals AI) | 26.6% | 32.6% | Hy4 preview |
| Legal Research Bench (Vals AI) | 41.4% | 45.2% | Hy4 preview |
| Harvey's Legal Agent Benchmark (Vals AI) | Not measured | 8.8% | Only Hy4 preview measured |
| Excel Modeling Benchmark (Vals AI) | 57.2% | 58.1% | Hy4 preview |
| BioMysteryBench (Vals AI) | 67.8% | 69.3% | Hy4 preview |
| Terminal-Bench-Science (Vals AI) | 0.0% | 1.4% | Hy4 preview |
| Vals Science Index | 29.6% | Not measured | Only DeepSeek V4.1 Flash measured |
| APEX-Agents | 39.5% | Not measured | Only DeepSeek V4.1 Flash measured |
| Neither model has a result on the other 28 boards here | |||
ReasoningDirectional onlyPublic boards, shown as published, kept out of the index13 boards, none shared
| Benchmark | DeepSeek V4.1 Flash | Hy4 preview | Leader |
|---|---|---|---|
| LiveBench reasoning | 86.7% | Not measured | Only DeepSeek V4.1 Flash measured |
| MysteryMechanism (Vals AI) | 21.2% | Not measured | Only DeepSeek V4.1 Flash measured |
| Humanity's Last Exam (Artificial Analysis run) | 39.2% | Not measured | Only DeepSeek V4.1 Flash measured |
| GPQA Diamond (Epoch AI run) | 89.8% | Not measured | Only DeepSeek V4.1 Flash measured |
| ARC-AGI | 94.5% | Not measured | Only DeepSeek V4.1 Flash measured |
| ARC-AGI-2 | 72.9% | Not measured | Only DeepSeek V4.1 Flash measured |
| Chess puzzles | 48.0% | Not measured | Only DeepSeek V4.1 Flash measured |
| Mystery game puzzles | 43.0% | Not measured | Only DeepSeek V4.1 Flash measured |
| LMArena Text: hard prompts | 1483 | Not measured | Only DeepSeek V4.1 Flash measured |
| LMArena Text: expert | 1501 | Not measured | Only DeepSeek V4.1 Flash measured |
| LMArena Text: hard prompts in English | 1488 | Not measured | Only DeepSeek V4.1 Flash measured |
| LMArena Vision: diagrams | 1297 | Not measured | Only DeepSeek V4.1 Flash measured |
| LMArena Vision: homework | 1307 | Not measured | Only DeepSeek V4.1 Flash measured |
| Neither model has a result on the other 9 boards here | |||
MathLike-for-likePublic boards, shown as published, kept out of the index01
| Benchmark | DeepSeek V4.1 Flash | Hy4 preview | Leader |
|---|---|---|---|
| OTIS Mock AIME | 98.3% | Not measured | Only DeepSeek V4.1 Flash measured |
| FrontierMath, tiers 1 to 3 | 67.4% | Not measured | Only DeepSeek V4.1 Flash measured |
| FrontierMath, tier 4 | 26.8% | Not measured | Only DeepSeek V4.1 Flash measured |
| ProofBench | 54.0% | 75.0% | Hy4 preview |
| LMArena Text: math | 1478 | Not measured | Only DeepSeek V4.1 Flash measured |
| LMArena Text: mathematical work | 1485 | Not measured | Only DeepSeek V4.1 Flash measured |
| Neither model has a result on the other 7 boards here | |||
KnowledgeLike-for-likePublic boards, shown as published, kept out of the index02
| Benchmark | DeepSeek V4.1 Flash | Hy4 preview | Leader |
|---|---|---|---|
| LegalBench (Vals AI) | 83.3% | 83.8% | Hy4 preview |
| MedCode (Vals AI) | 41.2% | 43.3% | Hy4 preview |
| SimpleQA Verified | 42.9% | Not measured | Only DeepSeek V4.1 Flash measured |
| LMArena Text: science | 1469 | Not measured | Only DeepSeek V4.1 Flash measured |
| LMArena Text: medicine and healthcare | 1484 | Not measured | Only DeepSeek V4.1 Flash measured |
| LMArena Text: legal and government | 1443 | Not measured | Only DeepSeek V4.1 Flash measured |
| LMArena Text: business and finance | 1461 | Not measured | Only DeepSeek V4.1 Flash measured |
| Neither model has a result on the other 11 boards here | |||
Web designDirectional onlyPublic boards, shown as published, kept out of the index9 boards, none shared
| Benchmark | DeepSeek V4.1 Flash | Hy4 preview | Leader |
|---|---|---|---|
| Design Arena: websites | 1309 | Not measured | Only DeepSeek V4.1 Flash measured |
| Design Arena: UI components | 1315 | Not measured | Only DeepSeek V4.1 Flash measured |
| Design Arena: game dev | 1330 | Not measured | Only DeepSeek V4.1 Flash measured |
| Design Arena: data viz | 1290 | Not measured | Only DeepSeek V4.1 Flash measured |
| Design Arena: 3D | 1356 | Not measured | Only DeepSeek V4.1 Flash measured |
| Design Arena: all code categories | 1320 | Not measured | Only DeepSeek V4.1 Flash measured |
| Design Arena: SVG | 1244 | Not measured | Only DeepSeek V4.1 Flash measured |
| Design Arena: ASCII art | 1265 | Not measured | Only DeepSeek V4.1 Flash measured |
| LiveBench design | 79.2% | Not measured | Only DeepSeek V4.1 Flash measured |
GeneralLike-for-likePublic boards, shown as published, kept out of the index11
| Benchmark | DeepSeek V4.1 Flash | Hy4 preview | Leader |
|---|---|---|---|
| Vals Index | Not measured | 53.2% | Only Hy4 preview measured |
| LiveBench data analysis | 79.3% | Not measured | Only DeepSeek V4.1 Flash measured |
| Public Benefits Bench (Vals AI) | 64.3% | 68.6% | Hy4 preview |
| SAGE (Vals AI) | 47.9% | Not measured | Only DeepSeek V4.1 Flash measured |
| MedScribe (Vals AI) | 85.5% | 83.6% | DeepSeek V4.1 Flash |
| Long context reasoning (AA-LCR) | 84.0% | Not measured | Only DeepSeek V4.1 Flash measured |
| DTBench | 89.9% | Not measured | Only DeepSeek V4.1 Flash measured |
| LMCA | 47.0% | Not measured | Only DeepSeek V4.1 Flash measured |
| Surface Evolver Bench | 46.3% | Not measured | Only DeepSeek V4.1 Flash measured |
| Furniture assembly | 34.2% | Not measured | Only DeepSeek V4.1 Flash measured |
| LMArena Text | 1464 | Not measured | Only DeepSeek V4.1 Flash measured |
| LMArena Text: instruction following | 1473 | Not measured | Only DeepSeek V4.1 Flash measured |
| LMArena Text: creative writing | 1437 | Not measured | Only DeepSeek V4.1 Flash measured |
| LMArena Vision | 1278 | Not measured | Only DeepSeek V4.1 Flash measured |
| LMArena Text: English | 1473 | Not measured | Only DeepSeek V4.1 Flash measured |
| LMArena Text: non-English | 1449 | Not measured | Only DeepSeek V4.1 Flash measured |
| LMArena Text: Chinese | 1501 | Not measured | Only DeepSeek V4.1 Flash measured |
| LMArena Text: German | 1470 | Not measured | Only DeepSeek V4.1 Flash measured |
| LMArena Text: Japanese | 1432 | Not measured | Only DeepSeek V4.1 Flash measured |
| LMArena Text: Korean | 1436 | Not measured | Only DeepSeek V4.1 Flash measured |
| LMArena Text: French | 1465 | Not measured | Only DeepSeek V4.1 Flash measured |
| LMArena Text: Spanish | 1459 | Not measured | Only DeepSeek V4.1 Flash measured |
| LMArena Text: Russian | 1467 | Not measured | Only DeepSeek V4.1 Flash measured |
| LMArena Text: Polish | 1457 | Not measured | Only DeepSeek V4.1 Flash measured |
| LMArena Text: multi-turn | 1461 | Not measured | Only DeepSeek V4.1 Flash measured |
| LMArena Text: longer queries | 1473 | Not measured | Only DeepSeek V4.1 Flash measured |
| LMArena Text: writing and language | 1450 | Not measured | Only DeepSeek V4.1 Flash measured |
| LMArena Text: entertainment and media | 1445 | Not measured | Only DeepSeek V4.1 Flash measured |
| LMArena Vision: English | 1288 | Not measured | Only DeepSeek V4.1 Flash measured |
| LMArena Vision: Chinese | 1327 | Not measured | Only DeepSeek V4.1 Flash measured |
| LMArena Vision: reading text | 1282 | Not measured | Only DeepSeek V4.1 Flash measured |
| LMArena Vision: humor | 1260 | Not measured | Only DeepSeek V4.1 Flash measured |
| LMArena Vision: creative writing | 1275 | Not measured | Only DeepSeek V4.1 Flash measured |
| Neither model has a result on the other 13 boards here | |||
Image generationNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
| Benchmark | DeepSeek V4.1 Flash | Hy4 preview | Leader |
|---|---|---|---|
| Neither model has a result on 5 boards here | |||
What each workload costs
Three fixed token mixes at each model's listed price, so per-token rates become a bill.
Chat turn
1K input + 500 output tokens
DeepSeek V4.1 Flash$0.0006Fits in one request
Hy4 preview$0.0019Fits in one request
DeepSeek V4.1 Flash costs less
Repository review
50K input + 3K output tokens
DeepSeek V4.1 Flash$0.013Fits in one request
Hy4 preview$0.044Fits in one request
DeepSeek V4.1 Flash costs less
Agent loop
200K input + 10K output tokens
DeepSeek V4.1 Flash$0.050Fits in one request
Hy4 preview$0.173Fits in one request
DeepSeek V4.1 Flash costs less
Listed OpenRouter rates, without cache or batch discounts. A workload that does not fit a model's context window would need splitting.
Specification differences
The leading value in a row is marked; a row with a missing value names no leader. Our two rows are marked in red.
| Specification | DeepSeek V4.1 Flash | Tencent Hy4 preview |
|---|---|---|
| Maker | DeepSeek | Tencent |
| Weights | Open weights | Open weights |
| Released | September 10, 2026 | August 28, 2026 |
| Context window | 1.05M | 1.05M |
| Input $/Mtok | $0.21 | $0.7506 |
| Output $/Mtok | $0.78 | $2.2509 |
| Coding index | 62.5 | 44.3 |
| Independent measurements behind it | 5 | 3 |
| Cost per solved task | Not computable | Not computable |
| Source | OpenRouter listing | OpenRouter listing |
Questions
Is DeepSeek V4.1 Flash better than Hy4 preview for coding?
Pick DeepSeek V4.1 Flash for coding: it scores 62.5 on the Vibe Coding Index against 44.3 for Hy4 preview. Pick Hy4 preview for work measured by LMArena WebDev or Code migration (Vals AI), where it leads.
Which is cheaper, DeepSeek V4.1 Flash or Hy4 preview?
DeepSeek V4.1 Flash costs $0.21 per million input tokens and $0.78 per million output tokens; Hy4 preview costs $0.7506 and $2.2509. Cost per solved task is known for neither of them.
Which has the bigger context window, DeepSeek V4.1 Flash or Hy4 preview?
Neither: both take 1.05M tokens.
How many benchmarks is the DeepSeek V4.1 Flash vs Hy4 preview comparison based on?
22 benchmarks have a result for both models as of October 21, 2026. DeepSeek V4.1 Flash's coding index rests on 5 independent benchmarks and Hy4 preview's on 3. Where only one model has a result, the page says so instead of naming a winner.