Qwen3.7 Plus vs Hy4 preview
Hy4 preview leads Qwen3.7 Plus on coding, 44.3 to 33.2 on the Vibe Coding Index, and Qwen3.7 Plus costs 1.8x less per output token ($1.28 vs $2.2509 per million). Updated October 21, 2026; 7 benchmarks have a result for both.
Qwen3.7 Plus or Hy4 preview: which to pick
Pick Hy4 preview for coding: it scores 44.3 on the Vibe Coding Index against 33.2 for Qwen3.7 Plus.
Pick Qwen3.7 Plus for output tokens 1.8x cheaper ($1.28 vs $2.2509 per million).
- Shared results
- 7
- Qwen3.7 Plus only
- 68
- Hy4 preview only
- 17
- Like-for-like categories
- 2 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
Hy4 preview
Hy4 preview scores 44.3 against 33.2.
2 and 3 benchmarks behind the scores
Intelligence
AA Intelligence Index
No call
Only Qwen3.7 Plus has a AA Intelligence Index result, 25.2.
Directional only
Agentic work
AA Agentic Index
No call
Only Qwen3.7 Plus has a AA Agentic Index result, 17.5.
Directional only
Web design
Design Arena: websites
No call
Only Qwen3.7 Plus has a Design Arena: websites result, 1269.
Directional only
General
Vals Index
No call
Only Hy4 preview has a Vals Index result, 53.2%.
Directional only
Long documents
Listed context window
Hy4 preview
Hy4 preview takes 1.05M tokens against 1M.
Listed by OpenRouter
Everyday cost
1K input + 500 output tokens
Qwen3.7 Plus
Qwen3.7 Plus $0.0010, Hy4 preview $0.0019 per turn.
List prices
At a glance
- Coding index
- Qwen3.7 Plus 33.2 from 2 benchmarks, Hy4 preview 44.3 from 3 benchmarks.
- Widest category
- Coding, 22.6 points apart on average over 3 shared benchmarks.
- Widest single gap
- Code migration (Vals AI): Qwen3.7 Plus 12.9% against Hy4 preview 47.4%, 34.6 points apart on a 0 to 100 scale.
- Output price
- Qwen3.7 Plus $1.28 against Hy4 preview $2.2509 per million output tokens, 1.8x apart.
- Context window
- Qwen3.7 Plus takes 1M tokens, Hy4 preview 1.05M.
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 | Qwen3.7 Plus | Bars | Hy4 preview | Ties |
|---|---|---|---|---|
| Coding | 0 | 3 | ||
| Agentic work | 0 | 4 | ||
| All shared boards | 0 | 7 |
CodingLike-for-likeSix of these boards feed the Vibe Coding Index03
| Benchmark | Qwen3.7 Plus | Hy4 preview | Leader |
|---|---|---|---|
| LMArena WebDev | Not measured | 1631 | Only Hy4 preview measured |
| AA Coding Index | 55.9 | Not measured | Only Qwen3.7 Plus measured |
| Terminal-Bench 4 (Vals AI) | Not measured | 8.1% | Only Hy4 preview measured |
| Vibe Code Bench (Vals AI) | 46.4% | 77.5% | Hy4 preview |
| Code migration (Vals AI) | 12.9% | 47.4% | Hy4 preview |
| ProgramBench (Vals AI) | Not measured | 0.0% | Only Hy4 preview measured |
| IOI (Vals AI) | Not measured | 59.3% | Only Hy4 preview measured |
| Terminal-Bench 2.1 (Vals AI) | 52.8% | 55.1% | Hy4 preview |
| SkillsBench (Vals AI) | 54.3% | Not measured | Only Qwen3.7 Plus measured |
| CyberBench (Vals AI) | Not measured | 65.4% | Only Hy4 preview measured |
| SciCode | 46.1% | Not measured | Only Qwen3.7 Plus measured |
| Terminal-Bench Hard (Artificial Analysis run) | 47.0% | Not measured | Only Qwen3.7 Plus measured |
| FrontierCode | 10.2% | Not measured | Only Qwen3.7 Plus measured |
| LMArena Text: coding | 1473 | Not measured | Only Qwen3.7 Plus measured |
| LMArena Text: software and IT | 1473 | Not measured | Only Qwen3.7 Plus measured |
| LMArena WebDev: React | Not measured | 1653 | Only Hy4 preview measured |
| LMArena WebDev: HTML | Not measured | 1620 | Only Hy4 preview measured |
| Neither model has a result on the other 22 boards here | |||
IntelligenceDirectional onlyPublic boards, shown as published, kept out of the index2 boards, none shared
| Benchmark | Qwen3.7 Plus | Hy4 preview | Leader |
|---|---|---|---|
| AA Intelligence Index | 25.2 | Not measured | Only Qwen3.7 Plus measured |
| Epoch Capabilities Index | 147.0 | Not measured | Only Qwen3.7 Plus measured |
Agentic workLike-for-likePublic boards, shown as published, kept out of the index04
| Benchmark | Qwen3.7 Plus | Hy4 preview | Leader |
|---|---|---|---|
| AA Agentic Index | 17.5 | Not measured | Only Qwen3.7 Plus measured |
| SREBench (Vals AI) | Not measured | 2.3% | Only Hy4 preview measured |
| Finance Agent (Vals AI) | 38.2% | 55.1% | Hy4 preview |
| Tax Agent Bench (Vals AI) | 10.5% | 32.6% | Hy4 preview |
| Legal Research Bench (Vals AI) | 16.4% | 45.2% | Hy4 preview |
| Harvey's Legal Agent Benchmark (Vals AI) | Not measured | 8.8% | Only Hy4 preview measured |
| Excel Modeling Benchmark (Vals AI) | 49.3% | 58.1% | Hy4 preview |
| BioMysteryBench (Vals AI) | Not measured | 69.3% | Only Hy4 preview measured |
| Terminal-Bench-Science (Vals AI) | Not measured | 1.4% | Only Hy4 preview measured |
| τ²-Bench Telecom | 93.0% | Not measured | Only Qwen3.7 Plus measured |
| OSWorld 2.0 | 2.8% | Not measured | Only Qwen3.7 Plus measured |
| Neither model has a result on the other 30 boards here | |||
ReasoningDirectional onlyPublic boards, shown as published, kept out of the index9 boards, none shared
| Benchmark | Qwen3.7 Plus | Hy4 preview | Leader |
|---|---|---|---|
| GPQA Diamond (Artificial Analysis run) | 90.0% | Not measured | Only Qwen3.7 Plus measured |
| Humanity's Last Exam (Artificial Analysis run) | 35.6% | Not measured | Only Qwen3.7 Plus measured |
| GPQA Diamond (Epoch AI run) | 87.9% | Not measured | Only Qwen3.7 Plus measured |
| Mystery game puzzles | 17.0% | Not measured | Only Qwen3.7 Plus measured |
| LMArena Text: hard prompts | 1460 | Not measured | Only Qwen3.7 Plus measured |
| LMArena Text: expert | 1466 | Not measured | Only Qwen3.7 Plus measured |
| LMArena Text: hard prompts in English | 1460 | Not measured | Only Qwen3.7 Plus measured |
| LMArena Vision: diagrams | 1292 | Not measured | Only Qwen3.7 Plus measured |
| LMArena Vision: homework | 1295 | Not measured | Only Qwen3.7 Plus measured |
| Neither model has a result on the other 13 boards here | |||
MathDirectional onlyPublic boards, shown as published, kept out of the index5 boards, none shared
| Benchmark | Qwen3.7 Plus | Hy4 preview | Leader |
|---|---|---|---|
| OTIS Mock AIME | 93.3% | Not measured | Only Qwen3.7 Plus measured |
| FrontierMath, tiers 1 to 3 | 34.4% | Not measured | Only Qwen3.7 Plus measured |
| ProofBench | Not measured | 75.0% | Only Hy4 preview measured |
| LMArena Text: math | 1467 | Not measured | Only Qwen3.7 Plus measured |
| LMArena Text: mathematical work | 1470 | Not measured | Only Qwen3.7 Plus measured |
| Neither model has a result on the other 8 boards here | |||
KnowledgeDirectional onlyPublic boards, shown as published, kept out of the index7 boards, none shared
| Benchmark | Qwen3.7 Plus | Hy4 preview | Leader |
|---|---|---|---|
| LegalBench (Vals AI) | Not measured | 83.8% | Only Hy4 preview measured |
| MortgageTax (Vals AI) | 66.2% | Not measured | Only Qwen3.7 Plus measured |
| MedCode (Vals AI) | Not measured | 43.3% | Only Hy4 preview measured |
| LMArena Text: science | 1464 | Not measured | Only Qwen3.7 Plus measured |
| LMArena Text: medicine and healthcare | 1464 | Not measured | Only Qwen3.7 Plus measured |
| LMArena Text: legal and government | 1451 | Not measured | Only Qwen3.7 Plus measured |
| LMArena Text: business and finance | 1435 | Not measured | Only Qwen3.7 Plus measured |
| Neither model has a result on the other 11 boards here | |||
Web designDirectional onlyPublic boards, shown as published, kept out of the index8 boards, none shared
| Benchmark | Qwen3.7 Plus | Hy4 preview | Leader |
|---|---|---|---|
| Design Arena: websites | 1269 | Not measured | Only Qwen3.7 Plus measured |
| Design Arena: UI components | 1261 | Not measured | Only Qwen3.7 Plus measured |
| Design Arena: game dev | 1262 | Not measured | Only Qwen3.7 Plus measured |
| Design Arena: data viz | 1276 | Not measured | Only Qwen3.7 Plus measured |
| Design Arena: 3D | 1239 | Not measured | Only Qwen3.7 Plus measured |
| Design Arena: all code categories | 1269 | Not measured | Only Qwen3.7 Plus measured |
| Design Arena: SVG | 1208 | Not measured | Only Qwen3.7 Plus measured |
| Design Arena: ASCII art | 1143 | Not measured | Only Qwen3.7 Plus measured |
| Neither model has a result on the other 1 board here | |||
GeneralDirectional onlyPublic boards, shown as published, kept out of the index33 boards, none shared
| Benchmark | Qwen3.7 Plus | Hy4 preview | Leader |
|---|---|---|---|
| Vals Index | Not measured | 53.2% | Only Hy4 preview measured |
| Public Benefits Bench (Vals AI) | Not measured | 68.6% | Only Hy4 preview measured |
| SAGE (Vals AI) | 39.3% | Not measured | Only Qwen3.7 Plus measured |
| MedScribe (Vals AI) | Not measured | 83.6% | Only Hy4 preview measured |
| Vals Multimodal Index | 53.9% | Not measured | Only Qwen3.7 Plus measured |
| IFBench | 78.0% | Not measured | Only Qwen3.7 Plus measured |
| Long context reasoning (AA-LCR) | 73.0% | Not measured | Only Qwen3.7 Plus measured |
| DTBench | 84.0% | Not measured | Only Qwen3.7 Plus measured |
| LMCA | 37.6% | Not measured | Only Qwen3.7 Plus measured |
| LMArena Text | 1454 | Not measured | Only Qwen3.7 Plus measured |
| LMArena Text: instruction following | 1439 | Not measured | Only Qwen3.7 Plus measured |
| LMArena Text: creative writing | 1438 | Not measured | Only Qwen3.7 Plus measured |
| LMArena Vision | 1278 | Not measured | Only Qwen3.7 Plus measured |
| LMArena Text: English | 1458 | Not measured | Only Qwen3.7 Plus measured |
| LMArena Text: non-English | 1445 | Not measured | Only Qwen3.7 Plus measured |
| LMArena Text: Chinese | 1509 | Not measured | Only Qwen3.7 Plus measured |
| LMArena Text: German | 1471 | Not measured | Only Qwen3.7 Plus measured |
| LMArena Text: Japanese | 1412 | Not measured | Only Qwen3.7 Plus measured |
| LMArena Text: Korean | 1413 | Not measured | Only Qwen3.7 Plus measured |
| LMArena Text: French | 1472 | Not measured | Only Qwen3.7 Plus measured |
| LMArena Text: Spanish | 1457 | Not measured | Only Qwen3.7 Plus measured |
| LMArena Text: Russian | 1457 | Not measured | Only Qwen3.7 Plus measured |
| LMArena Text: Polish | 1449 | Not measured | Only Qwen3.7 Plus measured |
| LMArena Text: multi-turn | 1459 | Not measured | Only Qwen3.7 Plus measured |
| LMArena Text: longer queries | 1454 | Not measured | Only Qwen3.7 Plus measured |
| LMArena Text: writing and language | 1437 | Not measured | Only Qwen3.7 Plus measured |
| LMArena Text: entertainment and media | 1422 | Not measured | Only Qwen3.7 Plus measured |
| LMArena Vision: English | 1280 | Not measured | Only Qwen3.7 Plus measured |
| LMArena Vision: Chinese | 1332 | Not measured | Only Qwen3.7 Plus measured |
| LMArena Vision: reading text | 1288 | Not measured | Only Qwen3.7 Plus measured |
| LMArena Vision: humor | 1261 | Not measured | Only Qwen3.7 Plus measured |
| LMArena Vision: creative writing | 1258 | Not measured | Only Qwen3.7 Plus measured |
| LMArena Document | 1444 | Not measured | Only Qwen3.7 Plus 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 | Qwen3.7 Plus | 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
Qwen3.7 Plus costs less
Repository review
50K input + 3K output tokens
Qwen3.7 Plus costs less
Agent loop
200K input + 10K output tokens
Qwen3.7 Plus 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 | Qwen3.7 Plus | Tencent Hy4 preview |
|---|---|---|
| Maker | Qwen | Tencent |
| Weights | No public weights listed | Open weights |
| Released | June 3, 2026 | August 28, 2026 |
| Context window | 1M | 1.05M |
| Input $/Mtok | $0.32 | $0.7506 |
| Output $/Mtok | $1.28 | $2.2509 |
| Coding index | 33.2 | 44.3 |
| Independent measurements behind it | 2 | 3 |
| Cost per solved task | Not computable | Not computable |
| Source | OpenRouter listing | OpenRouter listing |
Questions
Is Qwen3.7 Plus better than Hy4 preview for coding?
Pick Hy4 preview for coding: it scores 44.3 on the Vibe Coding Index against 33.2 for Qwen3.7 Plus. Pick Qwen3.7 Plus for output tokens 1.8x cheaper ($1.28 vs $2.2509 per million).
Which is cheaper, Qwen3.7 Plus or Hy4 preview?
Qwen3.7 Plus costs $0.32 per million input tokens and $1.28 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, Qwen3.7 Plus or Hy4 preview?
Hy4 preview, with 1.05M tokens against 1M.
How many benchmarks is the Qwen3.7 Plus vs Hy4 preview comparison based on?
7 benchmarks have a result for both models as of October 21, 2026. Qwen3.7 Plus's coding index rests on 2 independent benchmarks and Hy4 preview's on 3. Where only one model has a result, the page says so instead of naming a winner.