o3 vs Hy3 preview
Updated October 21, 2026. 31 benchmarks have a result for both. Coding is ranked on the Vibe Coding Index; every other board is shown as published.
Decision reading
Pick o3 for coding: it scores 70.7 on the Vibe Coding Index against 22.4 for Hy3 preview.
Pick Hy3 preview when output tokens cost 13.3x less, or when the context window is larger, 262K against 200K tokens.
- Shared results
- 31
- o3 only
- 66
- Hy3 preview only
- 5
- Like-for-like categories
- 7 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
o3
o3 scores 70.7 against 22.4.
2 and 1 benchmarks behind the scores
Intelligence
AA Intelligence Index
Hy3 preview
Hy3 preview leads, 25.3 against 20.2.
Same board, both measured
Agentic work
AA Agentic Index
No call
Only Hy3 preview has a AA Agentic Index result, 24.1.
Directional only
Reasoning
GPQA Diamond
No call
Only o3 has a GPQA Diamond result, 84.1%.
Directional only
Math
AIME
No call
Only o3 has a AIME result, 85.3%.
Directional only
Knowledge
MMLU Pro
No call
Only o3 has a MMLU Pro result, 85.6%.
Directional only
Web design
Design Arena: websites
No call
Only o3 has a Design Arena: websites result, 1036.
Directional only
Long documents
Listed context window
Hy3 preview
Hy3 preview takes 262K tokens against 200K.
Listed by OpenRouter
Everyday cost
1K input + 500 output tokens
Hy3 preview
o3 $0.0060, Hy3 preview $0.0005 per turn.
List prices
Cost per solved task
One SWE-bench Verified task, solved
No call
Only o3 has the number, $0.571.
Directional only
At a glance
- Coding index
- o3 70.7 from 2 benchmarks, Hy3 preview 22.4 from 1 benchmark.
- Widest category
- Agentic work, 12.0 points apart on average over 1 shared benchmark.
- Widest single gap
- τ²-Bench Telecom: o3 80.7% against Hy3 preview 92.7%, 12.0 points apart on a 0 to 100 scale.
- Output price
- o3 $8.00 against Hy3 preview $0.60 per million output tokens, 13.3x apart.
- Context window
- o3 takes 200K 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 | o3 | Bars | Hy3 preview | Ties |
|---|---|---|---|---|
| Coding | 1 | 2 | ||
| Intelligence | 0 | 1 | ||
| Agentic work | 0 | 1 | ||
| Reasoning | 0 | 5 | ||
| Math | 1 | 1 | ||
| Knowledge | 2 | 2 | ||
| General | 11 | 4 | ||
| All shared boards | 15 | 16 |
CodingLike-for-likeFour of these boards feed the Vibe Coding Index12
| Benchmark | o3 | Hy3 preview | Leader |
|---|---|---|---|
| Aider polyglot | 81.3% | Not measured | Only o3 measured |
| SWE-bench Verified | 58.4% | Not measured | Only o3 measured |
| SWE-bench Multimodal | 36.0% | Not measured | Only o3 measured |
| LMArena WebDev | Not measured | 1357 | Only Hy3 preview measured |
| AA Coding Index | Not measured | 58.8 | Only Hy3 preview measured |
| LiveCodeBench (Vals AI) | 83.9% | Not measured | Only o3 measured |
| LiveCodeBench (Artificial Analysis run) | 80.8% | Not measured | Only o3 measured |
| Terminal-Bench Hard (Artificial Analysis run) | 37.1% | 34.1% | o3 |
| GSO-Bench | 8.8% | Not measured | Only o3 measured |
| WeirdML | 52.4% | Not measured | Only o3 measured |
| ALE-Bench | 933.5 | Not measured | Only o3 measured |
| LMArena Text: coding | 1407 | 1421 | Hy3 preview |
| LMArena Text: software and IT | 1415 | 1423 | Hy3 preview |
| LMArena WebDev: React | Not measured | 1349 | Only Hy3 preview measured |
| LMArena WebDev: HTML | Not measured | 1356 | Only Hy3 preview measured |
| Neither model has a result on the other 23 boards here | |||
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index01
| Benchmark | o3 | Hy3 preview | Leader |
|---|---|---|---|
| AA Intelligence Index | 20.2 | 25.3 | Hy3 preview |
| Epoch Capabilities Index | 146.7 | Not measured | Only o3 measured |
Agentic workLike-for-likePublic boards, shown as published, kept out of the index01
| Benchmark | o3 | Hy3 preview | Leader |
|---|---|---|---|
| AA Agentic Index | Not measured | 24.1 | Only Hy3 preview measured |
| τ²-Bench Telecom | 80.7% | 92.7% | Hy3 preview |
| METR time horizons | 65.4% | Not measured | Only o3 measured |
| GDPval | 30.8% | Not measured | Only o3 measured |
| DeepResearch Bench | 45.2% | Not measured | Only o3 measured |
| OSWorld | 23.0% | Not measured | Only o3 measured |
| Neither model has a result on the other 33 boards here | |||
ReasoningLike-for-likePublic boards, shown as published, kept out of the index05
| Benchmark | o3 | Hy3 preview | Leader |
|---|---|---|---|
| GPQA Diamond | 84.1% | Not measured | Only o3 measured |
| MMMU Pro (Vals AI) | 80.4% | Not measured | Only o3 measured |
| GPQA Diamond (Artificial Analysis run) | 82.7% | 86.7% | Hy3 preview |
| Humanity's Last Exam (Artificial Analysis run) | 20.1% | 27.8% | Hy3 preview |
| GPQA Diamond (Epoch AI run) | 81.8% | Not measured | Only o3 measured |
| ARC-AGI | 60.8% | Not measured | Only o3 measured |
| ARC-AGI-2 | 6.5% | Not measured | Only o3 measured |
| SimpleBench | 53.1% | Not measured | Only o3 measured |
| Chess puzzles | 37.0% | Not measured | Only o3 measured |
| Mystery game puzzles | 29.0% | Not measured | Only o3 measured |
| VPCT | 52.0% | Not measured | Only o3 measured |
| LMArena Text: hard prompts | 1402 | 1413 | Hy3 preview |
| LMArena Text: expert | 1402 | 1422 | Hy3 preview |
| LMArena Text: hard prompts in English | 1411 | 1416 | Hy3 preview |
| LMArena Vision: diagrams | 1220 | Not measured | Only o3 measured |
| LMArena Vision: homework | 1234 | Not measured | Only o3 measured |
| Neither model has a result on the other 5 boards here | |||
MathLike-for-likePublic boards, shown as published, kept out of the index11
| Benchmark | o3 | Hy3 preview | Leader |
|---|---|---|---|
| AIME | 85.3% | Not measured | Only o3 measured |
| MATH 500 (Vals AI) | 94.6% | Not measured | Only o3 measured |
| MGSM (Vals AI) | 91.8% | Not measured | Only o3 measured |
| AA Math Index | 88.3 | Not measured | Only o3 measured |
| AIME 2025 (Artificial Analysis run) | 88.3% | Not measured | Only o3 measured |
| OTIS Mock AIME | 84.4% | Not measured | Only o3 measured |
| FrontierMath, tiers 1 to 3 | 33.3% | Not measured | Only o3 measured |
| MATH level 5 | 97.8% | Not measured | Only o3 measured |
| LMArena Text: math | 1426 | 1424 | o3 |
| LMArena Text: mathematical work | 1423 | 1442 | Hy3 preview |
| Neither model has a result on the other 3 boards here | |||
KnowledgeLike-for-likePublic boards, shown as published, kept out of the index22
| Benchmark | o3 | Hy3 preview | Leader |
|---|---|---|---|
| MMLU Pro | 85.6% | Not measured | Only o3 measured |
| LegalBench (Vals AI) | 83.8% | Not measured | Only o3 measured |
| CorpFin (Vals AI) | 59.7% | Not measured | Only o3 measured |
| TaxEval (Vals AI) | 74.6% | Not measured | Only o3 measured |
| MortgageTax (Vals AI) | 65.7% | Not measured | Only o3 measured |
| MedQA (Vals AI) | 96.1% | Not measured | Only o3 measured |
| MedCode (Vals AI) | 47.3% | Not measured | Only o3 measured |
| MMLU Pro (Artificial Analysis run) | 85.3% | Not measured | Only o3 measured |
| SimpleQA Verified | 49.4% | Not measured | Only o3 measured |
| GeoBench | 74.0% | Not measured | Only o3 measured |
| LMArena Text: science | 1429 | 1416 | o3 |
| LMArena Text: medicine and healthcare | 1448 | 1405 | o3 |
| LMArena Text: legal and government | 1428 | 1430 | Hy3 preview |
| LMArena Text: business and finance | 1393 | 1397 | Hy3 preview |
| Neither model has a result on the other 3 boards here | |||
Web designDirectional onlyPublic boards, shown as published, kept out of the index5 boards, none shared
| Benchmark | o3 | Hy3 preview | Leader |
|---|---|---|---|
| Design Arena: websites | 1036 | Not measured | Only o3 measured |
| Design Arena: UI components | 1016 | Not measured | Only o3 measured |
| Design Arena: game dev | 1034 | Not measured | Only o3 measured |
| Design Arena: data viz | 1200 | Not measured | Only o3 measured |
| Design Arena: all code categories | 1021 | Not measured | Only o3 measured |
| Neither model has a result on the other 4 boards here | |||
GeneralLike-for-likePublic boards, shown as published, kept out of the index114
| Benchmark | o3 | Hy3 preview | Leader |
|---|---|---|---|
| SAGE (Vals AI) | 41.8% | Not measured | Only o3 measured |
| MedScribe (Vals AI) | 76.7% | Not measured | Only o3 measured |
| IFBench | 71.4% | 63.1% | o3 |
| Long context reasoning (AA-LCR) | 74.7% | 64.7% | o3 |
| DTBench | 84.8% | Not measured | Only o3 measured |
| LMCA | 39.7% | Not measured | Only o3 measured |
| Fiction.LiveBench | 88.9% | Not measured | Only o3 measured |
| Creative writing (Lech Mazur) | 83.9% | Not measured | Only o3 measured |
| CL-bench | 17.8% | Not measured | Only o3 measured |
| LMArena Text | 1410 | 1401 | o3 |
| LMArena Text: instruction following | 1368 | 1383 | Hy3 preview |
| LMArena Text: creative writing | 1359 | 1339 | o3 |
| LMArena Vision | 1214 | Not measured | Only o3 measured |
| LMArena Text: English | 1417 | 1409 | o3 |
| LMArena Text: non-English | 1401 | 1390 | o3 |
| LMArena Text: Chinese | 1437 | 1435 | o3 |
| LMArena Text: German | 1420 | Not measured | Only o3 measured |
| LMArena Text: Japanese | 1403 | Not measured | Only o3 measured |
| LMArena Text: Korean | 1370 | Not measured | Only o3 measured |
| LMArena Text: French | 1430 | 1425 | o3 |
| LMArena Text: Spanish | 1395 | 1399 | Hy3 preview |
| LMArena Text: Russian | 1406 | 1399 | o3 |
| LMArena Text: Polish | 1438 | Not measured | Only o3 measured |
| LMArena Text: multi-turn | 1405 | 1411 | Hy3 preview |
| LMArena Text: longer queries | 1372 | 1402 | Hy3 preview |
| LMArena Text: writing and language | 1372 | 1370 | o3 |
| LMArena Text: entertainment and media | 1373 | 1352 | o3 |
| LMArena Vision: English | 1221 | Not measured | Only o3 measured |
| LMArena Vision: Chinese | 1237 | Not measured | Only o3 measured |
| LMArena Vision: reading text | 1218 | Not measured | Only o3 measured |
| LMArena Vision: humor | 1222 | Not measured | Only o3 measured |
| LMArena Vision: captioning | 1210 | Not measured | Only o3 measured |
| LMArena Vision: recognizing things | 1234 | Not measured | Only o3 measured |
| LMArena Vision: creative writing | 1201 | Not measured | Only o3 measured |
| Neither model has a result on the other 12 boards here | |||
Image generationNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
| Benchmark | o3 | 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 o3 | Tencent Hy3 preview |
|---|---|---|
| Maker | OpenAI | Tencent |
| Weights | No public weights listed | Open weights |
| Released | April 16, 2025 | April 22, 2026 |
| Context window | 200K | 262K |
| Input $/Mtok | $2.00 | $0.18 |
| Output $/Mtok | $8.00 | $0.60 |
| Coding index | 70.7 | 22.4 |
| Independent measurements behind it | 2 | 1 |
| Cost per solved task | $0.571 | Not computable |
| Source | OpenRouter listing | OpenRouter listing |
Questions
Is o3 better than Hy3 preview for coding?
Pick o3 for coding: it scores 70.7 on the Vibe Coding Index against 22.4 for Hy3 preview. Pick Hy3 preview when output tokens cost 13.3x less, or when the context window is larger, 262K against 200K tokens.
Which is cheaper, o3 or Hy3 preview?
o3 costs $2.00 per million input tokens and $8.00 per million output tokens; Hy3 preview costs $0.18 and $0.60. Cost per solved task is known for o3 only.
Which has the bigger context window, o3 or Hy3 preview?
Hy3 preview, with 262K tokens against 200K.
How many benchmarks is the o3 vs Hy3 preview comparison based on?
31 benchmarks have a result for both models as of October 21, 2026. o3'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.