GPT-4o vs Hy3 preview
Hy3 preview costs 16.7x less per output token than GPT-4o ($0.60 vs $10.00 per million). Updated October 22, 2026; 6 benchmarks have a result for both.
GPT-4o or Hy3 preview: which to pick
Neither model has a coding score yet, so the two cannot be ranked on coding.
Pick GPT-4o for work measured by Terminal-Bench Hard (Artificial Analysis run) or τ²-Bench Telecom, where it leads.
- Shared results
- 6
- GPT-4o only
- 21
- Hy3 preview only
- 30
- Like-for-like categories
- 4 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
No call
Neither has a coding score yet.
One side not measured
Intelligence
AA Intelligence Index
No call
Only Hy3 preview has a AA Intelligence Index result, 25.3.
Directional only
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 GPT-4o has a Design Arena: websites result, 831.
Directional only
Long documents
Listed context window
Hy3 preview
Hy3 preview takes 262K tokens against 128K.
Listed by OpenRouter
Everyday cost
1K input + 500 output tokens
Hy3 preview
GPT-4o $0.0075, Hy3 preview $0.0005 per turn.
List prices
At a glance
- Coding index
- GPT-4o not benchmarked, Hy3 preview not benchmarked.
- Widest category
- Agentic work, 67.5 points apart on average over 1 shared benchmark.
- Widest single gap
- τ²-Bench Telecom: GPT-4o 25.1% against Hy3 preview 92.7%, 67.5 points apart on a 0 to 100 scale.
- Output price
- GPT-4o $10.00 against Hy3 preview $0.60 per million output tokens, 16.7x apart.
- Context window
- GPT-4o takes 128K 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 | GPT-4o | Bars | Hy3 preview | Ties |
|---|---|---|---|---|
| Coding | 0 | 1 | ||
| Agentic work | 0 | 1 | ||
| Reasoning | 0 | 2 | ||
| General | 0 | 2 | ||
| All shared boards | 0 | 6 |
CodingLike-for-likeSix of these boards feed the Vibe Coding Index01
| Benchmark | GPT-4o | Hy3 preview | Leader |
|---|---|---|---|
| SWE-bench Verified | 38.8% | Not measured | Only GPT-4o measured |
| SWE-bench Lite | 38.0% | Not measured | Only GPT-4o measured |
| SWE-bench (full test set) | 18.8% | Not measured | Only GPT-4o measured |
| SWE-bench Multimodal | 30.4% | Not measured | Only GPT-4o measured |
| LMArena WebDev | Not measured | 1357 | Only Hy3 preview measured |
| AA Coding Index | Not measured | 58.8 | Only Hy3 preview measured |
| LiveCodeBench (Artificial Analysis run) | 30.9% | Not measured | Only GPT-4o measured |
| Terminal-Bench Hard (Artificial Analysis run) | 8.3% | 34.1% | Hy3 preview |
| LMArena Text: coding | Not measured | 1421 | Only Hy3 preview measured |
| LMArena Text: software and IT | Not measured | 1423 | Only Hy3 preview measured |
| LMArena WebDev: React | Not measured | 1349 | Only Hy3 preview measured |
| LMArena WebDev: HTML | Not measured | 1356 | Only Hy3 preview measured |
| SciCode | 1.5% | Not measured | Only GPT-4o measured |
| Neither model has a result on the other 26 boards here | |||
IntelligenceDirectional onlyPublic boards, shown as published, kept out of the index1 board, none shared
| Benchmark | GPT-4o | Hy3 preview | Leader |
|---|---|---|---|
| AA Intelligence Index | Not measured | 25.3 | Only Hy3 preview measured |
| Neither model has a result on the other 1 board here | |||
Agentic workLike-for-likePublic boards, shown as published, kept out of the index01
| Benchmark | GPT-4o | Hy3 preview | Leader |
|---|---|---|---|
| AA Agentic Index | Not measured | 24.1 | Only Hy3 preview measured |
| τ²-Bench Telecom | 25.1% | 92.7% | Hy3 preview |
| GDPval | 9.9% | Not measured | Only GPT-4o measured |
| Neither model has a result on the other 38 boards here | |||
ReasoningLike-for-likePublic boards, shown as published, kept out of the index02
| Benchmark | GPT-4o | Hy3 preview | Leader |
|---|---|---|---|
| GPQA Diamond (Artificial Analysis run) | 54.3% | 86.7% | Hy3 preview |
| Humanity's Last Exam (Artificial Analysis run) | 2.4% | 27.8% | Hy3 preview |
| ARC-AGI | 4.5% | Not measured | Only GPT-4o measured |
| ARC-AGI-2 | 0.0% | Not measured | Only GPT-4o measured |
| LMArena Text: hard prompts | Not measured | 1413 | Only Hy3 preview measured |
| LMArena Text: expert | Not measured | 1422 | Only Hy3 preview measured |
| LMArena Text: hard prompts in English | Not measured | 1416 | Only Hy3 preview measured |
| Neither model has a result on the other 15 boards here | |||
MathDirectional onlyPublic boards, shown as published, kept out of the index4 boards, none shared
| Benchmark | GPT-4o | Hy3 preview | Leader |
|---|---|---|---|
| AA Math Index | 6.0 | Not measured | Only GPT-4o measured |
| AIME 2025 (Artificial Analysis run) | 6.0% | Not measured | Only GPT-4o measured |
| LMArena Text: math | Not measured | 1424 | Only Hy3 preview measured |
| LMArena Text: mathematical work | Not measured | 1442 | Only Hy3 preview measured |
| Neither model has a result on the other 9 boards here | |||
KnowledgeDirectional onlyPublic boards, shown as published, kept out of the index6 boards, none shared
| Benchmark | GPT-4o | Hy3 preview | Leader |
|---|---|---|---|
| MMLU Pro (Artificial Analysis run) | 74.8% | Not measured | Only GPT-4o measured |
| MMLU | 88.1% | Not measured | Only GPT-4o measured |
| LMArena Text: science | Not measured | 1416 | Only Hy3 preview measured |
| LMArena Text: medicine and healthcare | Not measured | 1405 | Only Hy3 preview measured |
| LMArena Text: legal and government | Not measured | 1430 | Only Hy3 preview measured |
| LMArena Text: business and finance | Not measured | 1397 | Only Hy3 preview measured |
| Neither model has a result on the other 12 boards here | |||
Web designDirectional onlyPublic boards, shown as published, kept out of the index6 boards, none shared
| Benchmark | GPT-4o | Hy3 preview | Leader |
|---|---|---|---|
| Design Arena: websites | 831 | Not measured | Only GPT-4o measured |
| Design Arena: UI components | 882 | Not measured | Only GPT-4o measured |
| Design Arena: game dev | 906 | Not measured | Only GPT-4o measured |
| Design Arena: data viz | 858 | Not measured | Only GPT-4o measured |
| Design Arena: 3D | 868 | Not measured | Only GPT-4o measured |
| Design Arena: all code categories | 860 | Not measured | Only GPT-4o measured |
| Neither model has a result on the other 3 boards here | |||
GeneralLike-for-likePublic boards, shown as published, kept out of the index02
| Benchmark | GPT-4o | Hy3 preview | Leader |
|---|---|---|---|
| IFBench | 34.3% | 63.1% | Hy3 preview |
| Long context reasoning (AA-LCR) | 49.3% | 64.7% | Hy3 preview |
| DTBench | 64.5% | Not measured | Only GPT-4o measured |
| LMCA | 16.6% | Not measured | Only GPT-4o measured |
| LMArena Text | Not measured | 1401 | Only Hy3 preview measured |
| LMArena Text: instruction following | Not measured | 1383 | Only Hy3 preview measured |
| LMArena Text: creative writing | Not measured | 1339 | Only Hy3 preview measured |
| LMArena Text: English | Not measured | 1409 | Only Hy3 preview measured |
| LMArena Text: non-English | Not measured | 1390 | Only Hy3 preview measured |
| LMArena Text: Chinese | Not measured | 1435 | Only Hy3 preview measured |
| LMArena Text: French | Not measured | 1425 | Only Hy3 preview measured |
| LMArena Text: Spanish | Not measured | 1399 | Only Hy3 preview measured |
| LMArena Text: Russian | Not measured | 1399 | Only Hy3 preview measured |
| LMArena Text: multi-turn | Not measured | 1411 | Only Hy3 preview measured |
| LMArena Text: longer queries | Not measured | 1402 | Only Hy3 preview measured |
| LMArena Text: writing and language | Not measured | 1370 | Only Hy3 preview measured |
| LMArena Text: entertainment and media | Not measured | 1352 | Only Hy3 preview measured |
| Neither model has a result on the other 29 boards here | |||
Image generationNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
| Benchmark | GPT-4o | 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
Hy3 preview costs less
Repository review
50K input + 3K output tokens
Hy3 preview costs less
Agent loop
200K input + 10K output tokens
Hy3 preview 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 | OpenAI GPT-4o | Tencent Hy3 preview |
|---|---|---|
| Maker | OpenAI | Tencent |
| Weights | No public weights listed | Open weights |
| Released | May 13, 2024 | April 22, 2026 |
| Context window | 128K | 262K |
| Input $/Mtok | $2.50 | $0.18 |
| Output $/Mtok | $10.00 | $0.60 |
| Coding index | Not benchmarked | Not benchmarked |
| Independent measurements behind it | 0 | 1 |
| Cost per solved task | Not computable | Not computable |
| Source | OpenRouter listing | OpenRouter listing |
Questions
Is GPT-4o better than Hy3 preview for coding?
Neither model has a coding score yet, so the two cannot be ranked on coding. Pick GPT-4o for work measured by Terminal-Bench Hard (Artificial Analysis run) or τ²-Bench Telecom, where it leads.
Which is cheaper, GPT-4o or Hy3 preview?
GPT-4o costs $2.50 per million input tokens and $10.00 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, GPT-4o or Hy3 preview?
Hy3 preview, with 262K tokens against 128K.
How many benchmarks is the GPT-4o vs Hy3 preview comparison based on?
6 benchmarks have a result for both models as of October 22, 2026. GPT-4o's coding index rests on 0 independent benchmarks and Hy3 preview's on 1. Where only one model has a result, the page says so instead of naming a winner.