Model comparison
Hy4 preview vs Llama 3.2 90B
Hy4 preview is the stronger model overall, scoring 45.3 to 27.5 on the Noometry Index.
Last verified . 0 shared benchmarks.
Tencent
MetaSummary
- The widest gap is in math, where Hy4 preview leads 55.7 to 11.1.
Side by side
| Hy4 preview | Llama 3.2 90B | |
|---|---|---|
| Provider | Tencent | Meta |
| Noometry Index | 45.3 | 27.5 |
| Released | 2026-08-28 | 2024-09-24 |
| Weights | Open | Open |
| Context window | 1.05M | — |
| Max output | 64K | — |
| Input $ / M tokens | $0.83 | — |
| Output $ / M tokens | $2.50 | — |
| Results tracked | 3 | 9 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Not comparable
Hy4 preview: 51.6 (#38), Llama 3.2 90B: —
| Benchmark | Hy4 preview | Llama 3.2 90B |
|---|---|---|
| LMArena WebDev | 1632 | — |
Agentic & Tool Use Not comparable
Hy4 preview: —, Llama 3.2 90B: 30.0 (#80)
| Benchmark | Hy4 preview | Llama 3.2 90B |
|---|---|---|
| BALROG | — | 27.3% |
Reasoning Hy4 preview leads
Hy4 preview: 31.9 (#79), Llama 3.2 90B: 21.7 (#217)
| Benchmark | Hy4 preview | Llama 3.2 90B |
|---|---|---|
| NYT Connections (extended) | 68.2% | — |
| EnigmaEval | — | 0.4% |
| Epoch Capabilities Index | — | 125.5 |
Math Hy4 preview leads
Hy4 preview: 55.7 (#42), Llama 3.2 90B: 11.1 (#308)
| Benchmark | Hy4 preview | Llama 3.2 90B |
|---|---|---|
| OTIS Mock AIME 2024-2025 | — | 2.6% |
| ProofBench | 75% | — |
| MATH Level 5 | — | 39.4% |
Knowledge Not comparable
Hy4 preview: —, Llama 3.2 90B: 21.7 (#274)
| Benchmark | Hy4 preview | Llama 3.2 90B |
|---|---|---|
| GPQA Diamond | — | 41% |
| MMLU | — | 80.3% |
Multimodal Not comparable
Hy4 preview: —, Llama 3.2 90B: 25.4 (#124)
| Benchmark | Hy4 preview | Llama 3.2 90B |
|---|---|---|
| LMArena Vision | — | 1000 |
| GeoBench | — | 52% |
Frequently asked questions
Is Hy4 preview better than Llama 3.2 90B?
Hy4 preview is the stronger model overall, scoring 45.3 to 27.5 on the Noometry Index.
How many benchmarks do Hy4 preview and Llama 3.2 90B share?
0 benchmarks have published results for both models. Hy4 preview has 3 scored results on Noometry and Llama 3.2 90B has 9.