Hy4 preview vs Pareto 26.10 Preview
Hy4 preview costs 1.3x less per output token than Pareto 26.10 Preview ($2.501 vs $3.20 per million). Updated November 14, 2026; 1 benchmarks have a result for both.
Hy4 preview or Pareto 26.10 Preview: which to pick
Only Hy4 preview has a coding score, 44.3 on the Vibe Coding Index; Pareto 26.10 Preview has none, so the two cannot be ranked on coding.
Pick Pareto 26.10 Preview for work measured by GameBench 2, where it leads.
- Shared results
- 1
- Hy4 preview only
- 24
- Pareto 26.10 Preview only
- 0
- Like-for-like categories
- 1 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
Only Hy4 preview has a coding score, so no call is made.
One side not measured
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
Hy4 preview $0.0021, Pareto 26.10 Preview $0.0024 per turn.
List prices
At a glance
- Coding index
- Hy4 preview 44.3 from 3 benchmarks, Pareto 26.10 Preview not benchmarked.
- Widest category
- Coding, 27.5 points apart on average over 1 shared benchmark.
- Widest single gap
- GameBench 2: Hy4 preview 60.8 against Pareto 26.10 Preview 88.3, 27.5 points apart on a 0 to 100 scale.
- Output price
- Hy4 preview $2.501 against Pareto 26.10 Preview $3.20 per million output tokens, 1.3x 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 | Hy4 preview | Bars | Pareto 26.10 Preview | Ties |
|---|---|---|---|---|
| Coding | 0 | 1 | ||
| All shared boards | 0 | 1 |
CodingLike-for-likeSix of these boards feed the Vibe Coding Index01
| Benchmark | Hy4 preview | Pareto 26.10 Preview | Leader |
|---|---|---|---|
| LMArena WebDev | 1632 | Not measured | Only Hy4 preview measured |
| Terminal-Bench 4 (Vals AI) | 8.1% | Not measured | Only Hy4 preview measured |
| Vibe Code Bench (Vals AI) | 77.5% | Not measured | Only Hy4 preview measured |
| Code migration (Vals AI) | 47.4% | Not measured | Only Hy4 preview measured |
| ProgramBench (Vals AI) | 0.0% | Not measured | Only Hy4 preview measured |
| IOI (Vals AI) | 59.3% | Not measured | Only Hy4 preview measured |
| Terminal-Bench 2.1 (Vals AI) | 55.1% | Not measured | Only Hy4 preview measured |
| CyberBench (Vals AI) | 65.4% | Not measured | Only Hy4 preview measured |
| LMArena WebDev: React | 1653 | Not measured | Only Hy4 preview measured |
| LMArena WebDev: HTML | 1625 | Not measured | Only Hy4 preview measured |
| GameBench 2 | 60.8 | 88.3 | Pareto 26.10 Preview |
| Neither model has a result on the other 41 boards here | |||
IntelligenceNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
| Benchmark | Hy4 preview | Pareto 26.10 Preview | Leader |
|---|---|---|---|
| Neither model has a result on 2 boards here | |||
Agentic workDirectional onlyPublic boards, shown as published, kept out of the index8 boards, none shared
| Benchmark | Hy4 preview | Pareto 26.10 Preview | Leader |
|---|---|---|---|
| SREBench (Vals AI) | 2.3% | Not measured | Only Hy4 preview measured |
| Finance Agent (Vals AI) | 55.1% | Not measured | Only Hy4 preview measured |
| Tax Agent Bench (Vals AI) | 32.6% | Not measured | Only Hy4 preview measured |
| Legal Research Bench (Vals AI) | 45.2% | Not measured | Only Hy4 preview measured |
| Harvey's Legal Agent Benchmark (Vals AI) | 8.8% | Not measured | Only Hy4 preview measured |
| Excel Modeling Benchmark (Vals AI) | 58.1% | Not measured | Only Hy4 preview measured |
| BioMysteryBench (Vals AI) | 69.3% | Not measured | Only Hy4 preview measured |
| Terminal-Bench-Science (Vals AI) | 1.4% | Not measured | Only Hy4 preview measured |
| Neither model has a result on the other 36 boards here | |||
ReasoningNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
| Benchmark | Hy4 preview | Pareto 26.10 Preview | Leader |
|---|---|---|---|
| Neither model has a result on 26 boards here | |||
MathDirectional onlyPublic boards, shown as published, kept out of the index1 board, none shared
| Benchmark | Hy4 preview | Pareto 26.10 Preview | Leader |
|---|---|---|---|
| ProofBench | 75.0% | Not measured | Only Hy4 preview measured |
| Neither model has a result on the other 13 boards here | |||
KnowledgeDirectional onlyPublic boards, shown as published, kept out of the index2 boards, none shared
| Benchmark | Hy4 preview | Pareto 26.10 Preview | Leader |
|---|---|---|---|
| LegalBench (Vals AI) | 83.8% | Not measured | Only Hy4 preview measured |
| MedCode (Vals AI) | 43.3% | Not measured | Only Hy4 preview measured |
| Neither model has a result on the other 19 boards here | |||
Web designNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
| Benchmark | Hy4 preview | Pareto 26.10 Preview | Leader |
|---|---|---|---|
| Neither model has a result on 9 boards here | |||
GeneralDirectional onlyPublic boards, shown as published, kept out of the index3 boards, none shared
| Benchmark | Hy4 preview | Pareto 26.10 Preview | Leader |
|---|---|---|---|
| Vals Index | 53.2% | Not measured | Only Hy4 preview measured |
| Public Benefits Bench (Vals AI) | 68.6% | Not measured | Only Hy4 preview measured |
| MedScribe (Vals AI) | 83.6% | Not measured | Only Hy4 preview measured |
| Neither model has a result on the other 46 boards here | |||
Image generationNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
| Benchmark | Hy4 preview | Pareto 26.10 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
Hy4 preview costs less
Agent loop
200K input + 10K output tokens
Hy4 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 | Tencent Hy4 preview | Pareto 26.10 Preview |
|---|---|---|
| Maker | Tencent | Unbiased |
| Weights | Open weights | No public weights listed |
| Released | August 28, 2026 | October 1, 2026 |
| Context window | 1.05M | 1.05M |
| Input $/Mtok | $0.834 | $0.80 |
| Output $/Mtok | $2.501 | $3.20 |
| Coding index | 44.3 | Not benchmarked |
| Independent measurements behind it | 3 | 0 |
| Cost per solved task | Not computable | Not computable |
| Source | OpenRouter listing | OpenRouter listing |
Questions
Is Hy4 preview better than Pareto 26.10 Preview for coding?
Only Hy4 preview has a coding score, 44.3 on the Vibe Coding Index; Pareto 26.10 Preview has none, so the two cannot be ranked on coding. Pick Pareto 26.10 Preview for work measured by GameBench 2, where it leads.
Which is cheaper, Hy4 preview or Pareto 26.10 Preview?
Hy4 preview costs $0.834 per million input tokens and $2.501 per million output tokens; Pareto 26.10 Preview costs $0.80 and $3.20. Cost per solved task is known for neither of them.
Which has the bigger context window, Hy4 preview or Pareto 26.10 Preview?
Neither: both take 1.05M tokens.
How many benchmarks is the Hy4 preview vs Pareto 26.10 Preview comparison based on?
1 benchmarks have a result for both models as of November 14, 2026. Hy4 preview's coding index rests on 3 independent benchmarks and Pareto 26.10 Preview's on 0. Where only one model has a result, the page says so instead of naming a winner.