Qwen 3.8 Max vs Grok 4.5: Benchmark Comparison
Qwen 3.8 Max has the higher score on 14 of 25 shared benchmarks; Grok 4.5 leads on 10.
The largest observed score gap is 42.44 pts on CyberBench v1.1 , where Grok 4.5 leads.
Reported ±1 standard-error ranges overlap on 10 of 25 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 | Grok 4.5 | Gap | Reported uncertainty |
|---|---|---|---|---|
| Vals Index | 48.27% ±1.34 | 44.72% ±1.17 | 3.55 pts | Reported ±1 SE ranges do not overlap |
| Harvey's Legal Agent Benchmark | 10.42% ±2.29 | 12.92% ±2.53 | 2.50 pts | Reported ±1 SE ranges overlap |
| Legal Research Bench | 47.60% ±3.47 | 37.98% ±3.37 | 9.61 pts | Reported ±1 SE ranges do not overlap |
| LegalBench | 83.61% ±0.42 | 85.97% ±0.41 | 2.36 pts | Reported ±1 SE ranges do not overlap |
| EMB | 60.07% ±3.35 | 52.91% ±3.11 | 7.16 pts | Reported ±1 SE ranges do not overlap |
| Finance Agent (v2) | 50.59% ±0.58 | 48.35% ±0.46 | 2.24 pts | Reported ±1 SE ranges do not overlap |
| MortgageTax | 63.99% ±0.94 | 61.80% ±0.97 | 2.19 pts | Reported ±1 SE ranges do not overlap |
| Tax Agent Bench | 65.96% ±3.18 | 64.32% ±3.16 | 1.64 pts | Reported ±1 SE ranges overlap |
| TaxEval v2 | 75.55% ±0.84 | 71.67% ±0.89 | 3.88 pts | Reported ±1 SE ranges do not overlap |
| MedCode | 40.67% ±2.03 | 43.29% ±2.31 | 2.62 pts | Reported ±1 SE ranges overlap |
| MedScribe | 84.95% ±2.00 | 86.88% ±1.94 | 1.94 pts | Reported ±1 SE ranges overlap |
| ProofBench v1.1 | 58.00% ±4.96 | 31.00% ±4.65 | 27.00 pts | Reported ±1 SE ranges do not overlap |
| GPQA Diamond | 93.69% ±1.22 | 92.93% ±1.29 | 0.76 pts | Reported ±1 SE ranges overlap |
| MMLU Pro | 88.60% ±0.31 | 89.22% ±0.31 | 0.61 pts | Reported ±1 SE ranges overlap |
| MMMU Pro | 88.03% ±0.78 | 61.79% ±1.17 | 26.24 pts | Reported ±1 SE ranges do not overlap |
| SAGE | 51.25% ±3.42 | 34.97% ±3.32 | 16.29 pts | Reported ±1 SE ranges do not overlap |
| Code Migration | 23.96% ±4.31 | 36.60% ±4.14 | 12.64 pts | Reported ±1 SE ranges do not overlap |
| IOI | 68.89% ±5.20 | 40.56% ±3.95 | 28.33 pts | Reported ±1 SE ranges do not overlap |
| LiveCodeBench | 87.85% ±0.95 | 87.35% ±0.96 | 0.50 pts | Reported ±1 SE ranges overlap |
| ProgramBench | 0.00% ±0.00 | 0.00% ±0.00 | 0.00 pts | Reported ±1 SE ranges overlap |
| SkillsBench | 42.01% ±4.43 | 66.03% ±4.60 | 24.02 pts | Reported ±1 SE ranges do not overlap |
| SWE-bench | 85.60% ±1.57 | 86.60% ±1.52 | 1.00 pts | Reported ±1 SE ranges overlap |
| Terminal-Bench 4.0 | 34.34% ±3.94 | 8.59% ±1.34 | 25.76 pts | Reported ±1 SE ranges do not overlap |
| Vibe Code Bench v1.1 | 64.70% ±5.36 | 69.00% ±4.48 | 4.31 pts | Reported ±1 SE ranges overlap |
| CyberBench v1.1 | 28.57% ±3.31 | 71.01% ±5.79 | 42.44 pts | Reported ±1 SE ranges do not overlap |
Performance by category
| Category | Qwen 3.8 Max average | Grok 4.5 average |
|---|---|---|
| Legal | 47.21% | 45.62% |
| Finance | 63.23% | 59.81% |
| Healthcare | 62.81% | 65.09% |
| Math | 58.00% | 31.00% |
| Academic | 90.11% | 81.31% |
| Education | 51.25% | 34.97% |
| Coding | 50.92% | 49.34% |
| Cyber | 28.57% | 71.01% |
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 | Grok 4.5 cost | Qwen 3.8 Max latency | Grok 4.5 latency |
|---|---|---|---|---|
| Vals Index | $4.96 | $2.46 | 1h41m | 15m44s |
| Harvey's Legal Agent Benchmark | $2.37 | $2.02 | 49m12s | 10m15s |
| Legal Research Bench | $2.49 | $1.09 | 1h01m | 15m54s |
| LegalBench | N/A | N/A | 36.37s | 67.88s |
| EMB | $2.66 | $1.55 | 57m47s | 18m28s |
| Finance Agent (v2) | $1.24 | $1.16 | 19m35s | 4m49s |
| MortgageTax | N/A | N/A | 47.50s | 117.46s |
| Tax Agent Bench | $1.55 | $0.47 | 40m36s | 4m44s |
| TaxEval v2 | N/A | N/A | 4m26s | 3m15s |
| MedCode | N/A | N/A | 6m18s | 59.64s |
| MedScribe | N/A | N/A | 4m11s | 11m45s |
| ProofBench v1.1 | $1.50 | $0.57 | 31m08s | 11m09s |
| GPQA Diamond | N/A | N/A | 4m27s | 3m38s |
| MMLU Pro | N/A | N/A | 63.58s | 5m58s |
| MMMU Pro | N/A | N/A | 43.85s | 11m55s |
| SAGE | N/A | N/A | 2m13s | 2m50s |
| Code Migration | $10.54 | $3.37 | 4h28m | 23m29s |
| IOI | $9.47 | $5.64 | 1h47m | 1h27m |
| LiveCodeBench | N/A | N/A | 5m38s | 9m49s |
| ProgramBench | $12.36 | $4.12 | 5h54m | 18m59s |
| SkillsBench | $0.52 | $0.58 | 15m48s | 2m16s |
| SWE-bench | $1.13 | $0.54 | 41m42s | 3m20s |
| Terminal-Bench 4.0 | $10.63 | $5.79 | 2h28m | 33m51s |
| Vibe Code Bench v1.1 | $8.24 | $4.10 | 2h49m | 13m50s |
| CyberBench v1.1 | $0.78 | $2.95 | 9m40s | 14m08s |
Results available only for Qwen 3.8 Max
- Vals RSI Index
- MysteryMechanism
- Terminal-Bench Science
- Vibe Code Bench 1-100
- Public Benefits Bench v1.1
Results available only for Grok 4.5
- SRE Bench
- Time Horizon Index: KSP