Claude Sonnet 5 vs Hy4 Preview: Benchmark Comparison
Claude Sonnet 5 has the higher score on 6 of 14 shared benchmarks; Hy4 Preview leads on 7.
The largest observed score gap is 14.33 pts on IOI , where Hy4 Preview leads.
Reported ±1 standard-error ranges overlap on 9 of 14 shared benchmarks with comparable uncertainty data. This is not a pairwise statistical significance test.
Shared benchmark results
Scores and reported standard errors come from the latest visible Vals benchmark version available for both models.
| Benchmark | Claude Sonnet 5 | Hy4 Preview | Gap | Reported uncertainty |
|---|---|---|---|---|
| Legal Research Bench | 41.83% ±3.43 | 45.19% ±3.46 | 3.37 pts | Reported ±1 SE ranges overlap |
| LegalBench | 83.92% ±0.46 | 83.76% ±0.41 | 0.16 pts | Reported ±1 SE ranges overlap |
| Tax Agent Bench | 29.37% ±3.25 | 32.62% ±3.38 | 3.24 pts | Reported ±1 SE ranges overlap |
| MedCode | 47.54% ±2.27 | 43.25% ±2.13 | 4.29 pts | Reported ±1 SE ranges overlap |
| MedScribe | 76.05% ±3.05 | 83.60% ±2.06 | 7.55 pts | Reported ±1 SE ranges do not overlap |
| ProofBench v1.1 | 77.00% ±4.23 | 75.00% ±4.35 | 2.00 pts | Reported ±1 SE ranges overlap |
| Terminal-Bench-Science | 5.71% ±2.79 | 1.43% ±1.43 | 4.29 pts | Reported ±1 SE ranges do not overlap |
| IOI | 45.00% ±2.75 | 59.33% ±4.60 | 14.33 pts | Reported ±1 SE ranges do not overlap |
| ProgramBench | 0.00% ±0.00 | 0.00% ±0.00 | 0.00 pts | Reported ±1 SE ranges overlap |
| Terminal-Bench 4.0 | 10.61% ±1.51 | 8.08% ±1.34 | 2.53 pts | Reported ±1 SE ranges overlap |
| Vibe Code Bench v1.1 | 81.33% ±3.05 | 77.48% ±4.04 | 3.85 pts | Reported ±1 SE ranges overlap |
| CyberBench v1.1 | 61.91% ±5.74 | 65.36% ±5.55 | 3.45 pts | Reported ±1 SE ranges overlap |
| SRE Bench | 0.38% ±0.38 | 2.29% ±0.93 | 1.91 pts | Reported ±1 SE ranges do not overlap |
| Public Benefits Bench v1.1 | 66.03% ±1.23 | 68.61% ±1.21 | 2.57 pts | Reported ±1 SE ranges do not overlap |
Performance by category
| Category | Claude Sonnet 5 average | Hy4 Preview average |
|---|---|---|
| Legal | 62.87% | 64.47% |
| Finance | 29.37% | 32.62% |
| Healthcare | 61.80% | 63.42% |
| Math | 77.00% | 75.00% |
| Science | 5.71% | 1.43% |
| Coding | 34.23% | 36.22% |
| Cyber | 31.14% | 33.82% |
| Social Mobility | 66.03% | 68.61% |
Cost and latency
Cost per test appears only where the benchmark reports it for both models. Latency is the measured completion time for that benchmark.
| Benchmark | Claude Sonnet 5 cost | Hy4 Preview cost | Claude Sonnet 5 latency | Hy4 Preview latency |
|---|---|---|---|---|
| Legal Research Bench | $2.72 | $0.85 | 25m46s | 1h04m |
| LegalBench | N/A | N/A | 4.90s | 90.56s |
| Tax Agent Bench | $1.79 | $0.59 | 19m19s | 41m04s |
| MedCode | N/A | N/A | 2m15s | 6m39s |
| MedScribe | N/A | N/A | 4m12s | 5m50s |
| ProofBench v1.1 | $1.37 | $0.34 | 14m58s | 22m54s |
| Terminal-Bench-Science | $30.30 | $2.14 | 3h06m | 3h48m |
| IOI | $12.89 | $1.35 | 59m44s | 1h01m |
| ProgramBench | $24.36 | $12.81 | 1h31m | 2h10m |
| Terminal-Bench 4.0 | $26.05 | $2.10 | 1h45m | 2h10m |
| Vibe Code Bench v1.1 | $25.39 | $2.18 | 1h07m | 41m33s |
| CyberBench v1.1 | $1.89 | $0.49 | 16m57s | 36m43s |
| SRE Bench | $18.49 | $4.90 | 3h53m | 1h40m |
| Public Benefits Bench v1.1 | $1.29 | $0.25 | 21m09s | 1h03m |
Results available only for Claude Sonnet 5
- MortgageTax
- TaxEval v2
- GPQA Diamond
- MMLU Pro
- MMMU Pro
- SAGE
- LiveCodeBench
- SkillsBench
- SWE-bench
- Vibe Code Bench 1-100
Results available only for Hy4 Preview
- Vals Finance Index
- Vals Index
- Vals Legal Index
- HLAB-Vals-1.1
- EMB v1.1
- Finance Agent v2.1
- BioMysteryBench
- Code Migration v1.1