Model comparison
DeepSeek-V2.5 (Sep 2024) vs Hy3
Hy3 is the stronger model overall, scoring 43.6 to 37.0 on the Noometry Index.
Last verified . 15 shared benchmarks.
DeepSeek
TencentSummary
- They share 15 benchmarks with published results for both. DeepSeek-V2.5 (Sep 2024) scores higher in 0 categories and Hy3 in 8 categories; 7 gaps are clear of the uncertainty.
- The widest gap is in coding, where Hy3 leads 46.6 to 31.3.
Side by side
| DeepSeek-V2.5 (Sep 2024) | Hy3 | |
|---|---|---|
| Provider | DeepSeek | Tencent |
| Noometry Index | 37.0 | 43.6 |
| Released | 2024-09-06 | 2026-07-06 |
| Weights | Open | Open |
| Context window | — | 262K |
| Max output | — | 128K |
| Input $ / M tokens | — | $0.13 |
| Output $ / M tokens | — | $0.53 |
| Results tracked | 22 | 17 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Hy3 leads
DeepSeek-V2.5 (Sep 2024): 31.3 (#274), Hy3: 46.6 (#60)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Hy3 |
|---|---|---|
| LMArena Coding | 1310 | 1465 |
| Aider Polyglot | 17.8% | — |
| LMArena WebDev | — | 1509 |
| BigCodeBench Instruct | 48.6% | — |
| BigCodeBench Complete | 53.2% | — |
| HumanEval+ | 83.5% | — |
| MBPP+ | 74.1% | — |
Reasoning Too close to call
DeepSeek-V2.5 (Sep 2024): 24.5 (#145), Hy3: 25.0 (#134)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Hy3 |
|---|---|---|
| LMArena Hard Prompts | 1289 | 1449 |
| NYT Connections (extended) | — | 41.2% |
Math Hy3 leads
DeepSeek-V2.5 (Sep 2024): 33.3 (#177), Hy3: 37.4 (#98)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Hy3 |
|---|---|---|
| LMArena Math | 1289 | 1475 |
Knowledge Hy3 leads
DeepSeek-V2.5 (Sep 2024): 34.2 (#190), Hy3: 40.2 (#108)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Hy3 |
|---|---|---|
| LMArena Expert | 1266 | 1461 |
Multilingual Hy3 leads
DeepSeek-V2.5 (Sep 2024): 42.7 (#188), Hy3: 53.7 (#60)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Hy3 |
|---|---|---|
| LMArena Non-English | 1273 | 1427 |
| LMArena Chinese | 1317 | 1496 |
| LMArena French | 1288 | 1461 |
| LMArena Korean | 1209 | 1390 |
| LMArena Russian | 1289 | 1435 |
| LMArena Spanish | 1248 | 1448 |
| LMArena German | 1258 | — |
| LMArena Japanese | 1228 | — |
Instruction Following Hy3 leads
DeepSeek-V2.5 (Sep 2024): 67.7 (#189), Hy3: 75.3 (#68)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Hy3 |
|---|---|---|
| LMArena Instruction Following | 1280 | 1425 |
Long Context Hy3 leads
DeepSeek-V2.5 (Sep 2024): 39.5 (#169), Hy3: 44.1 (#70)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Hy3 |
|---|---|---|
| LMArena Longer Query | 1301 | 1443 |
Writing & Preference Hy3 leads
DeepSeek-V2.5 (Sep 2024): 50.0 (#182), Hy3: 62.5 (#75)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Hy3 |
|---|---|---|
| LMArena Text | 1294 | 1442 |
| LMArena Creative Writing | 1285 | 1399 |
| LMArena Multi-Turn | 1297 | 1436 |
Frequently asked questions
Is DeepSeek-V2.5 (Sep 2024) better than Hy3?
Hy3 is the stronger model overall, scoring 43.6 to 37.0 on the Noometry Index.
Is DeepSeek-V2.5 (Sep 2024) or Hy3 better for coding?
Hy3 scores higher on coding benchmarks: 46.6 versus 31.3 in the Noometry coding category.
How many benchmarks do DeepSeek-V2.5 (Sep 2024) and Hy3 share?
15 benchmarks have published results for both models. DeepSeek-V2.5 (Sep 2024) has 22 scored results on Noometry and Hy3 has 17.