DeepSeek V4.1 Flash vs Hy3 preview
Updated October 21, 2026. 30 benchmarks have a result for both. Coding is ranked on the Vibe Coding Index; every other board is shown as published.
vs boards won of 30 shared
Decision reading
Pick DeepSeek V4.1 Flash for coding: it scores 77.2 on the Vibe Coding Index against 22.4 for Hy3 preview.
Pick Hy3 preview when the benchmark that matters is AA Intelligence Index, where it leads.
Model ADeepSeek V4.1 FlashDeepSeek77.2/1002 independent benchmarksNo cost per solved taskModel BHy3 previewTencent22.4/1001 independent benchmarkNo cost per solved task
- Shared results
- 30
- DeepSeek V4.1 Flash only
- 71
- Hy3 preview only
- 6
- Like-for-like categories
- 6 of 9
Shared results can support a head-to-head reading. Results only one model has describe coverage, not superiority.
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, four public coding benchmarks
DeepSeek V4.1 Flash
DeepSeek V4.1 Flash scores 77.2 against 22.4.
2 and 1 benchmarks behind the scores
Intelligence
AA Intelligence Index
Hy3 preview
Hy3 preview leads, 25.3 against 24.7.
Same board, both measured
Agentic work
AA Agentic Index
No call
Only Hy3 preview has a AA Agentic Index result, 24.1.
Directional only
Web design
Design Arena: websites
No call
Only DeepSeek V4.1 Flash has a Design Arena: websites result, 1309.
Directional only
Long documents
Listed context window
DeepSeek V4.1 Flash
DeepSeek V4.1 Flash takes 1.05M tokens against 262K.
Listed by OpenRouter
Everyday cost
1K input + 500 output tokens
DeepSeek V4.1 Flash
DeepSeek V4.1 Flash $0.0003, Hy3 preview $0.0005 per turn.
List prices
At a glance
- Coding index
- DeepSeek V4.1 Flash 77.2 from 2 benchmarks, Hy3 preview 22.4 from 1 benchmark.
- Widest category
- Coding, 29.1 points apart on average over 5 shared benchmarks.
- Widest single gap
- LMArena WebDev: React: DeepSeek V4.1 Flash 1628 against Hy3 preview 1349, 41.9 points apart on a 0 to 100 scale.
- Output price
- DeepSeek V4.1 Flash $0.525 against Hy3 preview $0.60 per million output tokens, 1.1x apart.
- Context window
- DeepSeek V4.1 Flash takes 1.05M tokens, Hy3 preview 262K.
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 | Hy3 preview | Ties |
|---|---|---|---|---|
| Coding | 5 | 0 | ||
| Intelligence | 0 | 1 | ||
| Reasoning | 4 | 0 | ||
| Math | 2 | 0 | ||
| Knowledge | 4 | 0 | ||
| General | 14 | 0 | ||
| All shared boards | 29 | 1 |
CodingLike-for-likeFour of these boards feed the Vibe Coding Index50
| Benchmark | DeepSeek V4.1 Flash | Hy3 preview | Leader |
|---|---|---|---|
| LiveBench coding | 90.2% | Not measured | Only DeepSeek V4.1 Flash measured |
| LMArena WebDev | 1618 | 1357 | DeepSeek V4.1 Flash |
| AA Coding Index | Not measured | 58.8 | Only Hy3 preview measured |
| Terminal-Bench 4 (Vals AI) | 19.7% | Not measured | Only DeepSeek V4.1 Flash measured |
| Vibe Code Bench (Vals AI) | 84.7% | Not measured | Only DeepSeek V4.1 Flash measured |
| Code migration (Vals AI) | 45.6% | Not measured | Only DeepSeek V4.1 Flash measured |
| ProgramBench (Vals AI) | 0.5% | Not measured | Only DeepSeek V4.1 Flash measured |
| IOI (Vals AI) | 40.3% | Not measured | Only DeepSeek V4.1 Flash measured |
| Terminal-Bench 2.1 (Vals AI) | 74.5% | Not measured | Only DeepSeek V4.1 Flash measured |
| 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% | Not measured | Only DeepSeek V4.1 Flash measured |
| SciCode | 51.9% | Not measured | Only DeepSeek V4.1 Flash measured |
| Terminal-Bench Hard (Artificial Analysis run) | Not measured | 34.1% | Only Hy3 preview 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 | 1421 | DeepSeek V4.1 Flash |
| LMArena Text: software and IT | 1496 | 1423 | DeepSeek V4.1 Flash |
| LMArena WebDev: React | 1628 | 1349 | DeepSeek V4.1 Flash |
| LMArena WebDev: HTML | 1612 | 1356 | DeepSeek V4.1 Flash |
| LMArena WebDev: from an image | 1606 | Not measured | Only DeepSeek V4.1 Flash measured |
| Neither model has a result on the other 16 boards here | |||
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index01
| Benchmark | DeepSeek V4.1 Flash | Hy3 preview | Leader |
|---|---|---|---|
| AA Intelligence Index | 24.7 | 25.3 | Hy3 preview |
| Epoch Capabilities Index | 155.9 | Not measured | Only DeepSeek V4.1 Flash measured |
Agentic workDirectional onlyPublic boards, shown as published, kept out of the index14 boards, none shared
| Benchmark | DeepSeek V4.1 Flash | Hy3 preview | Leader |
|---|---|---|---|
| AA Agentic Index | Not measured | 24.1 | Only Hy3 preview measured |
| 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% | Not measured | Only DeepSeek V4.1 Flash measured |
| Finance Agent (Vals AI) | 53.5% | Not measured | Only DeepSeek V4.1 Flash measured |
| Tax Agent Bench (Vals AI) | 26.6% | Not measured | Only DeepSeek V4.1 Flash measured |
| Legal Research Bench (Vals AI) | 41.4% | Not measured | Only DeepSeek V4.1 Flash measured |
| Excel Modeling Benchmark (Vals AI) | 57.2% | Not measured | Only DeepSeek V4.1 Flash measured |
| BioMysteryBench (Vals AI) | 67.8% | Not measured | Only DeepSeek V4.1 Flash measured |
| Terminal-Bench-Science (Vals AI) | 0.0% | Not measured | Only DeepSeek V4.1 Flash measured |
| Vals Science Index | 29.6% | Not measured | Only DeepSeek V4.1 Flash measured |
| τ²-Bench Telecom | Not measured | 92.7% | Only Hy3 preview measured |
| APEX-Agents | 39.5% | Not measured | Only DeepSeek V4.1 Flash measured |
| Neither model has a result on the other 25 boards here | |||
ReasoningLike-for-likePublic boards, shown as published, kept out of the index40
| Benchmark | DeepSeek V4.1 Flash | Hy3 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 |
| GPQA Diamond (Artificial Analysis run) | Not measured | 86.7% | Only Hy3 preview measured |
| Humanity's Last Exam (Artificial Analysis run) | 39.2% | 27.8% | DeepSeek V4.1 Flash |
| 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 | 1413 | DeepSeek V4.1 Flash |
| LMArena Text: expert | 1501 | 1422 | DeepSeek V4.1 Flash |
| LMArena Text: hard prompts in English | 1488 | 1416 | DeepSeek V4.1 Flash |
| 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 7 boards here | |||
MathLike-for-likePublic boards, shown as published, kept out of the index20
| Benchmark | DeepSeek V4.1 Flash | Hy3 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% | Not measured | Only DeepSeek V4.1 Flash measured |
| LMArena Text: math | 1478 | 1424 | DeepSeek V4.1 Flash |
| LMArena Text: mathematical work | 1485 | 1442 | DeepSeek V4.1 Flash |
| Neither model has a result on the other 7 boards here | |||
KnowledgeLike-for-likePublic boards, shown as published, kept out of the index40
| Benchmark | DeepSeek V4.1 Flash | Hy3 preview | Leader |
|---|---|---|---|
| LegalBench (Vals AI) | 83.3% | Not measured | Only DeepSeek V4.1 Flash measured |
| MedCode (Vals AI) | 41.2% | Not measured | Only DeepSeek V4.1 Flash measured |
| SimpleQA Verified | 42.9% | Not measured | Only DeepSeek V4.1 Flash measured |
| LMArena Text: science | 1469 | 1416 | DeepSeek V4.1 Flash |
| LMArena Text: medicine and healthcare | 1484 | 1405 | DeepSeek V4.1 Flash |
| LMArena Text: legal and government | 1443 | 1430 | DeepSeek V4.1 Flash |
| LMArena Text: business and finance | 1461 | 1397 | DeepSeek V4.1 Flash |
| Neither model has a result on the other 10 boards here | |||
Web designDirectional onlyPublic boards, shown as published, kept out of the index9 boards, none shared
| Benchmark | DeepSeek V4.1 Flash | Hy3 preview | Leader |
|---|---|---|---|
| Design Arena: websites | 1309 | Not measured | Only DeepSeek V4.1 Flash measured |
| Design Arena: UI components | 1314 | Not measured | Only DeepSeek V4.1 Flash measured |
| Design Arena: game dev | 1331 | 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 | 1243 | 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 index140
| Benchmark | DeepSeek V4.1 Flash | Hy3 preview | Leader |
|---|---|---|---|
| LiveBench data analysis | 79.3% | Not measured | Only DeepSeek V4.1 Flash measured |
| Public Benefits Bench (Vals AI) | 64.3% | Not measured | Only DeepSeek V4.1 Flash measured |
| SAGE (Vals AI) | 47.9% | Not measured | Only DeepSeek V4.1 Flash measured |
| MedScribe (Vals AI) | 85.5% | Not measured | Only DeepSeek V4.1 Flash measured |
| IFBench | Not measured | 63.1% | Only Hy3 preview measured |
| Long context reasoning (AA-LCR) | 84.0% | 64.7% | DeepSeek V4.1 Flash |
| 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 | 1401 | DeepSeek V4.1 Flash |
| LMArena Text: instruction following | 1473 | 1383 | DeepSeek V4.1 Flash |
| LMArena Text: creative writing | 1437 | 1339 | DeepSeek V4.1 Flash |
| LMArena Vision | 1278 | Not measured | Only DeepSeek V4.1 Flash measured |
| LMArena Text: English | 1473 | 1409 | DeepSeek V4.1 Flash |
| LMArena Text: non-English | 1449 | 1390 | DeepSeek V4.1 Flash |
| LMArena Text: Chinese | 1501 | 1435 | DeepSeek V4.1 Flash |
| 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 | 1425 | DeepSeek V4.1 Flash |
| LMArena Text: Spanish | 1459 | 1399 | DeepSeek V4.1 Flash |
| LMArena Text: Russian | 1467 | 1399 | DeepSeek V4.1 Flash |
| LMArena Text: Polish | 1457 | Not measured | Only DeepSeek V4.1 Flash measured |
| LMArena Text: multi-turn | 1461 | 1411 | DeepSeek V4.1 Flash |
| LMArena Text: longer queries | 1473 | 1402 | DeepSeek V4.1 Flash |
| LMArena Text: writing and language | 1450 | 1370 | DeepSeek V4.1 Flash |
| LMArena Text: entertainment and media | 1445 | 1352 | DeepSeek V4.1 Flash |
| 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 | Hy3 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$0.0003Fits in one request
Hy3 preview$0.0005Fits in one request
DeepSeek V4.1 Flash costs less
Repository review
50K input + 3K output tokens
DeepSeek V4.1 Flash$0.0040Fits in one request
Hy3 preview$0.011Fits in one request
DeepSeek V4.1 Flash costs less
Agent loop
200K input + 10K output tokens
DeepSeek V4.1 Flash$0.015Fits in one request
Hy3 preview$0.042Fits 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 Hy3 preview |
|---|---|---|
| Maker | DeepSeek | Tencent |
| Weights | Open weights | Open weights |
| Released | September 10, 2026 | April 22, 2026 |
| Context window | 1.05M | 262K |
| Input $/Mtok | $0.049 | $0.18 |
| Output $/Mtok | $0.525 | $0.60 |
| Coding index | 77.2 | 22.4 |
| Independent measurements behind it | 2 | 1 |
| Cost per solved task | Not computable | Not computable |
| Source | OpenRouter listing | OpenRouter listing |
Questions
Is DeepSeek V4.1 Flash better than Hy3 preview for coding?
Pick DeepSeek V4.1 Flash for coding: it scores 77.2 on the Vibe Coding Index against 22.4 for Hy3 preview. Pick Hy3 preview when the benchmark that matters is AA Intelligence Index, where it leads.
Which is cheaper, DeepSeek V4.1 Flash or Hy3 preview?
DeepSeek V4.1 Flash costs $0.049 per million input tokens and $0.525 per million output tokens; Hy3 preview costs $0.18 and $0.60. Cost per solved task is known for neither of them.
Which has the bigger context window, DeepSeek V4.1 Flash or Hy3 preview?
DeepSeek V4.1 Flash, with 1.05M tokens against 262K.
How many benchmarks is the DeepSeek V4.1 Flash vs Hy3 preview comparison based on?
30 benchmarks have a result for both models as of October 21, 2026. DeepSeek V4.1 Flash's coding index rests on 2 independent benchmarks and Hy3 preview's on 1. Where only one model has a result, the page says so instead of naming a winner.