Model comparison
Hy3 vs Qwen3.5 27B
Hy3 is the stronger model overall, scoring 44.2 to 41.9 on the Noometry Index.
Last verified . 19 shared benchmarks.
Tencent
Alibaba (Qwen)Summary
- They share 19 benchmarks with published results for both. Hy3 scores higher in 7 categories and Qwen3.5 27B in 1 category; 7 gaps are clear of the uncertainty.
- The widest gap is in coding, where Hy3 leads 46.8 to 38.9.
- The biggest single-benchmark swing is NYT Connections (extended): 41.2% for Hy3 and 47.9% for Qwen3.5 27B.
- Hy3 is cheaper at $0.0825 / $0.33 per million input/output tokens, against $0.30 / $2.40 for Qwen3.5 27B.
Side by side
| Hy3 | Qwen3.5 27B | |
|---|---|---|
| Provider | Tencent | Alibaba (Qwen) |
| Noometry Index | 44.2 | 41.9 |
| Released | 2026-07-06 | 2026-02-23 |
| Weights | Open | Open |
| Context window | 262K | 262K |
| Max output | 128K | 66K |
| Input $ / M tokens | $0.0825 | $0.30 |
| Output $ / M tokens | $0.33 | $2.40 |
| Results tracked | 19 | 28 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Hy3 leads
Hy3: 46.8 (#63), Qwen3.5 27B: 38.9 (#168)
| Benchmark | Hy3 | Qwen3.5 27B |
|---|---|---|
| LMArena WebDev | 1508 | 1358 |
| LMArena Coding | 1464 | 1427 |
| WeirdML | — | 39.5% |
| ALE-Bench | — | 349.45 |
Agentic & Tool Use Not comparable
Hy3: —, Qwen3.5 27B: —
| Benchmark | Hy3 | Qwen3.5 27B |
|---|---|---|
| Vending-Bench 2 | — | 201.98 |
Reasoning Qwen3.5 27B leads
Hy3: 26.1 (#136), Qwen3.5 27B: 27.5 (#117)
| Benchmark | Hy3 | Qwen3.5 27B |
|---|---|---|
| NYT Connections (extended) | 41.2% | 47.9% |
| LMArena Hard Prompts | 1447 | 1414 |
| Thematic Generalization | — | 45.5% |
| DTBench | — | 82.4% |
| LMCA | — | 34% |
Math Hy3 leads
Hy3: 40.1 (#93), Qwen3.5 27B: 38.8 (#127)
| Benchmark | Hy3 | Qwen3.5 27B |
|---|---|---|
| LMArena Math | 1475 | 1429 |
| MathArena Final-Answer Competitions | — | 56.7% |
Knowledge Hy3 leads
Hy3: 40.8 (#114), Qwen3.5 27B: 38.0 (#150)
| Benchmark | Hy3 | Qwen3.5 27B |
|---|---|---|
| LMArena Expert | 1460 | 1428 |
| Vectara Hallucination Rate | — | 12.1% |
Multimodal Not comparable
Hy3: —, Qwen3.5 27B: 39.4 (#59)
| Benchmark | Hy3 | Qwen3.5 27B |
|---|---|---|
| LMArena Vision | — | 1241 |
Multilingual Hy3 leads
Hy3: 53.5 (#65), Qwen3.5 27B: 50.8 (#115)
| Benchmark | Hy3 | Qwen3.5 27B |
|---|---|---|
| LMArena Non-English | 1426 | 1390 |
| LMArena Chinese | 1493 | 1478 |
| LMArena French | 1461 | 1410 |
| LMArena German | 1439 | 1393 |
| LMArena Japanese | 1392 | 1345 |
| LMArena Korean | 1395 | 1358 |
| LMArena Russian | 1432 | 1390 |
| LMArena Spanish | 1456 | 1407 |
Instruction Following Hy3 leads
Hy3: 75.1 (#70), Qwen3.5 27B: 73.5 (#119)
| Benchmark | Hy3 | Qwen3.5 27B |
|---|---|---|
| LMArena Instruction Following | 1426 | 1393 |
Long Context Too close to call
Hy3: 44.1 (#75), Qwen3.5 27B: 43.1 (#106)
| Benchmark | Hy3 | Qwen3.5 27B |
|---|---|---|
| LMArena Longer Query | 1442 | 1413 |
Writing & Preference Hy3 leads
Hy3: 62.2 (#81), Qwen3.5 27B: 59.3 (#111)
| Benchmark | Hy3 | Qwen3.5 27B |
|---|---|---|
| LMArena Text | 1439 | 1409 |
| LMArena Creative Writing | 1402 | 1362 |
| LMArena Multi-Turn | 1436 | 1410 |
Frequently asked questions
Is Hy3 better than Qwen3.5 27B?
Hy3 is the stronger model overall, scoring 44.2 to 41.9 on the Noometry Index.
Which is cheaper, Hy3 or Qwen3.5 27B?
Hy3 is cheaper. It lists at $0.0825 per million input tokens and $0.33 per million output tokens; Qwen3.5 27B lists at $0.30 and $2.40.
Is Hy3 or Qwen3.5 27B better for coding?
Hy3 scores higher on coding benchmarks: 46.8 versus 38.9 in the Noometry coding category.
Which has the bigger context window?
Both accept 262K tokens.
How many benchmarks do Hy3 and Qwen3.5 27B share?
19 benchmarks have published results for both models. Hy3 has 19 scored results on Noometry and Qwen3.5 27B has 28.