Gemini 3.6 Flash vs Hy4 preview
Gemini 3.6 Flash leads Hy4 preview on coding, 62.5 to 44.3 on the Vibe Coding Index, and Hy4 preview costs 1.5x less per output token ($2.501 vs $3.75 per million). Updated October 22, 2026; 20 benchmarks have a result for both.
Gemini 3.6 Flash or Hy4 preview: which to pick
Pick Gemini 3.6 Flash for coding: it scores 62.5 on the Vibe Coding Index against 44.3 for Hy4 preview.
Pick Hy4 preview for output tokens 1.5x cheaper ($2.501 vs $3.75 per million) and work measured by LMArena WebDev or Vibe Code Bench (Vals AI), where it leads.
- Shared results
- 20
- Gemini 3.6 Flash only
- 89
- Hy4 preview only
- 4
- 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
Gemini 3.6 Flash
Gemini 3.6 Flash scores 62.5 against 44.3.
4 and 3 benchmarks behind the scores
Intelligence
AA Intelligence Index
No call
Only Gemini 3.6 Flash has a AA Intelligence Index result, 34.0.
Directional only
Agentic work
AA Agentic Index
No call
Only Gemini 3.6 Flash has a AA Agentic Index result, 29.0.
Directional only
Reasoning
GPQA Diamond
No call
Only Gemini 3.6 Flash has a GPQA Diamond result, 93.4%.
Directional only
Knowledge
MMLU Pro
No call
Only Gemini 3.6 Flash has a MMLU Pro result, 89.3%.
Directional only
Web design
Design Arena: websites
No call
Only Gemini 3.6 Flash has a Design Arena: websites result, 1300.
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
Hy4 preview
Gemini 3.6 Flash $0.0026, Hy4 preview $0.0021 per turn.
List prices
Cost per solved task
One SWE-bench Verified task, solved
No call
Only Gemini 3.6 Flash has the number, $1.49.
Directional only
At a glance
- Coding index
- Gemini 3.6 Flash 62.5 from 4 benchmarks, Hy4 preview 44.3 from 3 benchmarks.
- Widest category
- Math, 39.0 points apart on average over 1 shared benchmark.
- Widest single gap
- ProofBench: Gemini 3.6 Flash 36.0% against Hy4 preview 75.0%, 39.0 points apart on a 0 to 100 scale.
- Output price
- Gemini 3.6 Flash $3.75 against Hy4 preview $2.501 per million output tokens, 1.5x 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 | Gemini 3.6 Flash | Bars | Hy4 preview | Ties |
|---|---|---|---|---|
| Coding | 1 | 7 | 1 tie | |
| Agentic work | 3 | 3 | ||
| Math | 0 | 1 | ||
| Knowledge | 2 | 0 | ||
| General | 0 | 2 | ||
| All shared boards | 6 | 13 | 1 tie |
CodingLike-for-likeSix of these boards feed the Vibe Coding Index17
| Benchmark | Gemini 3.6 Flash | Hy4 preview | Leader |
|---|---|---|---|
| LMArena WebDev | 1538 | 1631 | Hy4 preview |
| AA Coding Index | 69.2 | Not measured | Only Gemini 3.6 Flash measured |
| LiveCodeBench (Vals AI) | 88.1% | Not measured | Only Gemini 3.6 Flash measured |
| SWE-bench Verified (Vals AI) | 79.6% | Not measured | Only Gemini 3.6 Flash measured |
| Terminal-Bench 4 (Vals AI) | Not measured | 8.1% | Only Hy4 preview measured |
| Vibe Code Bench (Vals AI) | 64.0% | 77.5% | Hy4 preview |
| Code migration (Vals AI) | 30.9% | 47.4% | Hy4 preview |
| ProgramBench (Vals AI) | 0.0% | 0.0% | Tie |
| IOI (Vals AI) | 35.1% | 59.3% | Hy4 preview |
| Terminal-Bench 2.1 (Vals AI) | 73.8% | 55.1% | Gemini 3.6 Flash |
| CyberBench (Vals AI) | 45.4% | 65.4% | Hy4 preview |
| SciCode | 53.4% | Not measured | Only Gemini 3.6 Flash measured |
| FrontierCode | 34.4% | Not measured | Only Gemini 3.6 Flash measured |
| WeirdML | 56.1% | Not measured | Only Gemini 3.6 Flash measured |
| ALE-Bench | 715.5 | Not measured | Only Gemini 3.6 Flash measured |
| LMArena Text: coding | 1491 | Not measured | Only Gemini 3.6 Flash measured |
| LMArena Text: software and IT | 1492 | Not measured | Only Gemini 3.6 Flash measured |
| LMArena WebDev: React | 1528 | 1653 | Hy4 preview |
| LMArena WebDev: HTML | 1547 | 1620 | Hy4 preview |
| LMArena WebDev: from an image | 1545 | Not measured | Only Gemini 3.6 Flash measured |
| SciCode | 10.8% | Not measured | Only Gemini 3.6 Flash measured |
| Neither model has a result on the other 18 boards here | |||
IntelligenceDirectional onlyPublic boards, shown as published, kept out of the index2 boards, none shared
| Benchmark | Gemini 3.6 Flash | Hy4 preview | Leader |
|---|---|---|---|
| AA Intelligence Index | 34.0 | Not measured | Only Gemini 3.6 Flash measured |
| Epoch Capabilities Index | 154.5 | Not measured | Only Gemini 3.6 Flash measured |
Agentic workLike-for-likePublic boards, shown as published, kept out of the index33
| Benchmark | Gemini 3.6 Flash | Hy4 preview | Leader |
|---|---|---|---|
| AA Agentic Index | 29.0 | Not measured | Only Gemini 3.6 Flash measured |
| Design Arena: agents building web apps | 1197 | Not measured | Only Gemini 3.6 Flash measured |
| Design Arena: agents building full-stack apps | 1181 | Not measured | Only Gemini 3.6 Flash measured |
| Design Arena: agents building mobile apps | 1198 | Not measured | Only Gemini 3.6 Flash measured |
| Design Arena: agents building Android apps | 1204 | Not measured | Only Gemini 3.6 Flash measured |
| Design Arena: agents building games | 1180 | Not measured | Only Gemini 3.6 Flash measured |
| Design Arena: HTML slides | 1099 | Not measured | Only Gemini 3.6 Flash measured |
| Design Arena: PowerPoint slides | 1121 | Not measured | Only Gemini 3.6 Flash measured |
| Design Arena: spreadsheets | 1071 | Not measured | Only Gemini 3.6 Flash measured |
| SREBench (Vals AI) | Not measured | 2.3% | Only Hy4 preview measured |
| Finance Agent (Vals AI) | 56.3% | 55.1% | Gemini 3.6 Flash |
| Tax Agent Bench (Vals AI) | 18.0% | 32.6% | Hy4 preview |
| Legal Research Bench (Vals AI) | 25.0% | 45.2% | Hy4 preview |
| Harvey's Legal Agent Benchmark (Vals AI) | Not measured | 8.8% | Only Hy4 preview measured |
| Excel Modeling Benchmark (Vals AI) | 65.4% | 58.1% | Gemini 3.6 Flash |
| BioMysteryBench (Vals AI) | 58.5% | 69.3% | Hy4 preview |
| Terminal-Bench-Science (Vals AI) | 4.3% | 1.4% | Gemini 3.6 Flash |
| APEX-Agents | 46.9% | Not measured | Only Gemini 3.6 Flash measured |
| Neither model has a result on the other 23 boards here | |||
ReasoningDirectional onlyPublic boards, shown as published, kept out of the index14 boards, none shared
| Benchmark | Gemini 3.6 Flash | Hy4 preview | Leader |
|---|---|---|---|
| GPQA Diamond | 93.4% | Not measured | Only Gemini 3.6 Flash measured |
| MMMU Pro (Vals AI) | 88.4% | Not measured | Only Gemini 3.6 Flash measured |
| GPQA Diamond (Artificial Analysis run) | 92.8% | Not measured | Only Gemini 3.6 Flash measured |
| Humanity's Last Exam (Artificial Analysis run) | 40.8% | Not measured | Only Gemini 3.6 Flash measured |
| GPQA Diamond (Epoch AI run) | 94.1% | Not measured | Only Gemini 3.6 Flash measured |
| ARC-AGI | 91.2% | Not measured | Only Gemini 3.6 Flash measured |
| ARC-AGI-2 | 60.4% | Not measured | Only Gemini 3.6 Flash measured |
| Chess puzzles | 42.0% | Not measured | Only Gemini 3.6 Flash measured |
| Mystery game puzzles | 30.0% | Not measured | Only Gemini 3.6 Flash measured |
| LMArena Text: hard prompts | 1486 | Not measured | Only Gemini 3.6 Flash measured |
| LMArena Text: expert | 1490 | Not measured | Only Gemini 3.6 Flash measured |
| LMArena Text: hard prompts in English | 1484 | Not measured | Only Gemini 3.6 Flash measured |
| LMArena Vision: diagrams | 1314 | Not measured | Only Gemini 3.6 Flash measured |
| LMArena Vision: homework | 1314 | Not measured | Only Gemini 3.6 Flash measured |
| Neither model has a result on the other 8 boards here | |||
MathLike-for-likePublic boards, shown as published, kept out of the index01
| Benchmark | Gemini 3.6 Flash | Hy4 preview | Leader |
|---|---|---|---|
| OTIS Mock AIME | 94.2% | Not measured | Only Gemini 3.6 Flash measured |
| FrontierMath, tiers 1 to 3 | 58.9% | Not measured | Only Gemini 3.6 Flash measured |
| FrontierMath, tier 4 | 22.0% | Not measured | Only Gemini 3.6 Flash measured |
| ProofBench | 36.0% | 75.0% | Hy4 preview |
| LMArena Text: math | 1508 | Not measured | Only Gemini 3.6 Flash measured |
| LMArena Text: mathematical work | 1492 | Not measured | Only Gemini 3.6 Flash measured |
| Neither model has a result on the other 7 boards here | |||
KnowledgeLike-for-likePublic boards, shown as published, kept out of the index20
| Benchmark | Gemini 3.6 Flash | Hy4 preview | Leader |
|---|---|---|---|
| MMLU Pro | 89.3% | Not measured | Only Gemini 3.6 Flash measured |
| LegalBench (Vals AI) | 86.7% | 83.8% | Gemini 3.6 Flash |
| CorpFin (Vals AI) | 63.3% | Not measured | Only Gemini 3.6 Flash measured |
| TaxEval (Vals AI) | 74.9% | Not measured | Only Gemini 3.6 Flash measured |
| MortgageTax (Vals AI) | 67.6% | Not measured | Only Gemini 3.6 Flash measured |
| MedCode (Vals AI) | 53.1% | 43.3% | Gemini 3.6 Flash |
| SimpleQA Verified | 66.2% | Not measured | Only Gemini 3.6 Flash measured |
| LMArena Text: science | 1489 | Not measured | Only Gemini 3.6 Flash measured |
| LMArena Text: medicine and healthcare | 1469 | Not measured | Only Gemini 3.6 Flash measured |
| LMArena Text: legal and government | 1479 | Not measured | Only Gemini 3.6 Flash measured |
| LMArena Text: business and finance | 1464 | Not measured | Only Gemini 3.6 Flash measured |
| FACTS Benchmark Suite | 65.3% | Not measured | Only Gemini 3.6 Flash measured |
| Neither model has a result on the other 6 boards here | |||
Web designDirectional onlyPublic boards, shown as published, kept out of the index7 boards, none shared
| Benchmark | Gemini 3.6 Flash | Hy4 preview | Leader |
|---|---|---|---|
| Design Arena: websites | 1300 | Not measured | Only Gemini 3.6 Flash measured |
| Design Arena: UI components | 1294 | Not measured | Only Gemini 3.6 Flash measured |
| Design Arena: game dev | 1263 | Not measured | Only Gemini 3.6 Flash measured |
| Design Arena: data viz | 1289 | Not measured | Only Gemini 3.6 Flash measured |
| Design Arena: 3D | 1271 | Not measured | Only Gemini 3.6 Flash measured |
| Design Arena: all code categories | 1289 | Not measured | Only Gemini 3.6 Flash measured |
| Design Arena: ASCII art | 1281 | Not measured | Only Gemini 3.6 Flash measured |
| Neither model has a result on the other 2 boards here | |||
GeneralLike-for-likePublic boards, shown as published, kept out of the index02
| Benchmark | Gemini 3.6 Flash | Hy4 preview | Leader |
|---|---|---|---|
| Vals Index | Not measured | 53.2% | Only Hy4 preview measured |
| Public Benefits Bench (Vals AI) | 56.6% | 68.6% | Hy4 preview |
| SAGE (Vals AI) | 48.7% | Not measured | Only Gemini 3.6 Flash measured |
| MedScribe (Vals AI) | 79.7% | 83.6% | Hy4 preview |
| Vals Multimodal Index | 65.1% | Not measured | Only Gemini 3.6 Flash measured |
| Long context reasoning (AA-LCR) | 80.0% | Not measured | Only Gemini 3.6 Flash measured |
| DTBench | 95.5% | Not measured | Only Gemini 3.6 Flash measured |
| LMCA | 44.9% | Not measured | Only Gemini 3.6 Flash measured |
| Furniture assembly | 23.3% | Not measured | Only Gemini 3.6 Flash measured |
| LMArena Text | 1480 | Not measured | Only Gemini 3.6 Flash measured |
| LMArena Text: instruction following | 1467 | Not measured | Only Gemini 3.6 Flash measured |
| LMArena Text: creative writing | 1466 | Not measured | Only Gemini 3.6 Flash measured |
| LMArena Vision | 1299 | Not measured | Only Gemini 3.6 Flash measured |
| LMArena Text: English | 1482 | Not measured | Only Gemini 3.6 Flash measured |
| LMArena Text: non-English | 1470 | Not measured | Only Gemini 3.6 Flash measured |
| LMArena Text: Chinese | 1531 | Not measured | Only Gemini 3.6 Flash measured |
| LMArena Text: German | 1480 | Not measured | Only Gemini 3.6 Flash measured |
| LMArena Text: Japanese | 1475 | Not measured | Only Gemini 3.6 Flash measured |
| LMArena Text: Korean | 1429 | Not measured | Only Gemini 3.6 Flash measured |
| LMArena Text: French | 1503 | Not measured | Only Gemini 3.6 Flash measured |
| LMArena Text: Spanish | 1476 | Not measured | Only Gemini 3.6 Flash measured |
| LMArena Text: Russian | 1488 | Not measured | Only Gemini 3.6 Flash measured |
| LMArena Text: Polish | 1473 | Not measured | Only Gemini 3.6 Flash measured |
| LMArena Text: multi-turn | 1481 | Not measured | Only Gemini 3.6 Flash measured |
| LMArena Text: longer queries | 1475 | Not measured | Only Gemini 3.6 Flash measured |
| LMArena Text: writing and language | 1470 | Not measured | Only Gemini 3.6 Flash measured |
| LMArena Text: entertainment and media | 1456 | Not measured | Only Gemini 3.6 Flash measured |
| LMArena Vision: English | 1295 | Not measured | Only Gemini 3.6 Flash measured |
| LMArena Vision: Chinese | 1357 | Not measured | Only Gemini 3.6 Flash measured |
| LMArena Vision: reading text | 1312 | Not measured | Only Gemini 3.6 Flash measured |
| LMArena Vision: humor | 1289 | Not measured | Only Gemini 3.6 Flash measured |
| LMArena Vision: creative writing | 1278 | Not measured | Only Gemini 3.6 Flash measured |
| LMArena Document | 1456 | Not measured | Only Gemini 3.6 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 | Gemini 3.6 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
Hy4 preview costs less
Repository review
50K input + 3K output tokens
Gemini 3.6 Flash costs less
Agent loop
200K input + 10K output tokens
Gemini 3.6 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 | Google Gemini 3.6 Flash | Tencent Hy4 preview |
|---|---|---|
| Maker | Tencent | |
| Weights | No public weights listed | Open weights |
| Released | July 21, 2026 | August 28, 2026 |
| Context window | 1.05M | 1.05M |
| Input $/Mtok | $0.75 | $0.834 |
| Output $/Mtok | $3.75 | $2.501 |
| Coding index | 62.5 | 44.3 |
| Independent measurements behind it | 4 | 3 |
| Cost per solved task | $1.49 | Not computable |
| Source | OpenRouter listing | OpenRouter listing |
Questions
Is Gemini 3.6 Flash better than Hy4 preview for coding?
Pick Gemini 3.6 Flash for coding: it scores 62.5 on the Vibe Coding Index against 44.3 for Hy4 preview. Pick Hy4 preview for output tokens 1.5x cheaper ($2.501 vs $3.75 per million) and work measured by LMArena WebDev or Vibe Code Bench (Vals AI), where it leads.
Which is cheaper, Gemini 3.6 Flash or Hy4 preview?
Gemini 3.6 Flash costs $0.75 per million input tokens and $3.75 per million output tokens; Hy4 preview costs $0.834 and $2.501. Cost per solved task is known for Gemini 3.6 Flash only.
Which has the bigger context window, Gemini 3.6 Flash or Hy4 preview?
Neither: both take 1.05M tokens.
How many benchmarks is the Gemini 3.6 Flash vs Hy4 preview comparison based on?
20 benchmarks have a result for both models as of October 22, 2026. Gemini 3.6 Flash's coding index rests on 4 independent benchmarks and Hy4 preview's on 3. Where only one model has a result, the page says so instead of naming a winner.