Qwen 3.8 Max vs Hy4 Preview: Benchmark Comparison
Qwen 3.8 Max has the higher score on 8 of 18 shared benchmarks; Hy4 Preview leads on 9.
The largest observed score gap is 36.78 pts on CyberBench v1.1 , where Hy4 Preview leads.
Reported ±1 standard-error ranges overlap on 11 of 18 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 | Qwen 3.8 Max | Hy4 Preview | Gap | Reported uncertainty |
|---|---|---|---|---|
| Vals Index | 48.27% ±1.34 | 49.94% ±1.15 | 1.67 pts | Reported ±1 SE ranges overlap |
| Harvey's Legal Agent Benchmark | 10.42% ±2.29 | 9.17% ±2.00 | 1.25 pts | Reported ±1 SE ranges overlap |
| Legal Research Bench | 47.60% ±3.47 | 45.19% ±3.46 | 2.40 pts | Reported ±1 SE ranges overlap |
| LegalBench | 83.61% ±0.42 | 83.76% ±0.41 | 0.15 pts | Reported ±1 SE ranges overlap |
| EMB | 60.07% ±3.35 | 58.08% ±3.14 | 1.99 pts | Reported ±1 SE ranges overlap |
| Finance Agent (v2) | 50.59% ±0.58 | 55.06% ±0.31 | 4.47 pts | Reported ±1 SE ranges do not overlap |
| Tax Agent Bench | 65.96% ±3.18 | 63.71% ±3.26 | 2.25 pts | Reported ±1 SE ranges overlap |
| MedCode | 40.67% ±2.03 | 43.25% ±2.13 | 2.58 pts | Reported ±1 SE ranges overlap |
| MedScribe | 84.95% ±2.00 | 83.60% ±2.06 | 1.35 pts | Reported ±1 SE ranges overlap |
| ProofBench v1.1 | 58.00% ±4.96 | 75.00% ±4.35 | 17.00 pts | Reported ±1 SE ranges do not overlap |
| Terminal-Bench Science | 12.86% ±4.03 | 1.43% ±1.43 | 11.43 pts | Reported ±1 SE ranges do not overlap |
| Code Migration | 23.96% ±4.31 | 47.43% ±4.27 | 23.48 pts | Reported ±1 SE ranges do not overlap |
| IOI | 68.89% ±5.20 | 59.33% ±4.60 | 9.56 pts | Reported ±1 SE ranges overlap |
| ProgramBench | 0.00% ±0.00 | 0.00% ±0.00 | 0.00 pts | Reported ±1 SE ranges overlap |
| Terminal-Bench 4.0 | 34.34% ±3.94 | 8.08% ±1.34 | 26.26 pts | Reported ±1 SE ranges do not overlap |
| Vibe Code Bench v1.1 | 64.70% ±5.36 | 77.48% ±4.04 | 12.78 pts | Reported ±1 SE ranges do not overlap |
| CyberBench v1.1 | 28.57% ±3.31 | 65.36% ±5.55 | 36.78 pts | Reported ±1 SE ranges do not overlap |
| Public Benefits Bench v1.1 | 67.12% ±1.22 | 68.61% ±1.21 | 1.49 pts | Reported ±1 SE ranges overlap |
Performance by category
| Category | Qwen 3.8 Max average | Hy4 Preview average |
|---|---|---|
| Legal | 47.21% | 46.04% |
| Finance | 58.87% | 58.95% |
| Healthcare | 62.81% | 63.42% |
| Math | 58.00% | 75.00% |
| Science | 12.86% | 1.43% |
| Coding | 38.38% | 38.47% |
| Cyber | 28.57% | 65.36% |
| Social Mobility | 67.12% | 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 | Qwen 3.8 Max cost | Hy4 Preview cost | Qwen 3.8 Max latency | Hy4 Preview latency |
|---|---|---|---|---|
| Vals Index | $4.96 | $1.44 | 1h41m | 1h01m |
| Harvey's Legal Agent Benchmark | $2.37 | $0.98 | 49m12s | 29m25s |
| Legal Research Bench | $2.49 | $0.85 | 1h01m | 1h04m |
| LegalBench | N/A | N/A | 36.37s | 90.56s |
| EMB | $2.66 | $0.79 | 57m47s | 30m39s |
| Finance Agent (v2) | $1.24 | $0.59 | 19m35s | 23m45s |
| Tax Agent Bench | $1.55 | $0.59 | 40m36s | 41m04s |
| MedCode | N/A | N/A | 6m18s | 6m39s |
| MedScribe | N/A | N/A | 4m11s | 5m50s |
| ProofBench v1.1 | $1.50 | $0.34 | 31m08s | 22m54s |
| Terminal-Bench Science | $16.22 | $2.14 | 3h20m | 3h48m |
| Code Migration | $10.54 | $3.41 | 4h28m | 2h13m |
| IOI | $9.47 | $1.35 | 1h47m | 1h01m |
| ProgramBench | $12.36 | $12.81 | 5h54m | 2h10m |
| Terminal-Bench 4.0 | $10.63 | $2.10 | 2h28m | 2h10m |
| Vibe Code Bench v1.1 | $8.24 | $2.18 | 2h49m | 41m33s |
| CyberBench v1.1 | $0.78 | $0.49 | 9m40s | 36m43s |
| Public Benefits Bench v1.1 | $0.99 | $0.25 | 40m44s | 1h03m |
Results available only for Qwen 3.8 Max
- Vals RSI Index
- MortgageTax
- TaxEval v2
- MysteryMechanism
- GPQA Diamond
- MMLU Pro
- MMMU Pro
- SAGE
- LiveCodeBench
- SkillsBench
- SWE-bench
- Vibe Code Bench 1-100
Results available only for Hy4 Preview
- BioMysteryBench
- SRE Bench