Model comparison
Claude Sonnet 5 vs Grok 4.6: Benchmark Comparison
Claude Sonnet 5 has the higher score on 8 of 24 shared benchmarks; Grok 4.6 leads on 16.
The largest observed score gap is 26.00 pts on ProofBench v1.1 , where Claude Sonnet 5 leads.
Reported ±1 standard-error ranges overlap on 11 of 24 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 | Grok 4.6 | Gap | Reported uncertainty |
|---|---|---|---|---|
| Harvey's Legal Agent Benchmark | 5.00% ±1.65 | 15.83% ±2.65 | 10.83 pts | Reported ±1 SE ranges do not overlap |
| Legal Research Bench | 41.83% ±3.43 | 48.08% ±3.47 | 6.25 pts | Reported ±1 SE ranges overlap |
| LegalBench | 83.92% ±0.46 | 86.31% ±0.42 | 2.39 pts | Reported ±1 SE ranges do not overlap |
| EMB | 66.32% ±3.01 | 62.73% ±3.08 | 3.59 pts | Reported ±1 SE ranges overlap |
| Finance Agent (v2) | 53.91% ±0.52 | 53.68% ±0.67 | 0.23 pts | Reported ±1 SE ranges overlap |
| MortgageTax | 70.03% ±0.90 | 64.19% ±0.95 | 5.84 pts | Reported ±1 SE ranges do not overlap |
| Tax Agent Bench | 62.27% ±3.19 | 70.79% ±3.00 | 8.51 pts | Reported ±1 SE ranges do not overlap |
| TaxEval v2 | 75.63% ±0.84 | 71.10% ±0.90 | 4.54 pts | Reported ±1 SE ranges do not overlap |
| MedCode | 47.54% ±2.27 | 44.71% ±2.26 | 2.83 pts | Reported ±1 SE ranges overlap |
| MedScribe | 76.05% ±3.05 | 86.53% ±1.96 | 10.48 pts | Reported ±1 SE ranges do not overlap |
| ProofBench v1.1 | 77.00% ±4.23 | 51.00% ±5.02 | 26.00 pts | Reported ±1 SE ranges do not overlap |
| Terminal-Bench Science | 2.86% ±2.01 | 11.43% ±3.83 | 8.57 pts | Reported ±1 SE ranges do not overlap |
| GPQA Diamond | 88.89% ±2.22 | 94.70% ±1.13 | 5.81 pts | Reported ±1 SE ranges do not overlap |
| MMLU Pro | 87.55% ±0.37 | 89.40% ±0.30 | 1.86 pts | Reported ±1 SE ranges do not overlap |
| SAGE | 48.92% ±3.40 | 28.90% ±3.08 | 20.02 pts | Reported ±1 SE ranges do not overlap |
| Code Migration | 44.39% ±4.25 | 44.57% ±4.48 | 0.18 pts | Reported ±1 SE ranges overlap |
| IOI | 45.00% ±2.75 | 47.61% ±2.09 | 2.61 pts | Reported ±1 SE ranges overlap |
| LiveCodeBench | 82.43% ±1.09 | 88.22% ±0.94 | 5.80 pts | Reported ±1 SE ranges do not overlap |
| SkillsBench | 46.48% ±4.49 | 55.77% ±4.84 | 9.29 pts | Reported ±1 SE ranges overlap |
| SWE-bench | 79.60% ±1.80 | 95.60% ±0.92 | 16.00 pts | Reported ±1 SE ranges do not overlap |
| Vibe Code Bench 1-100 | 13.82% ±2.98 | 14.75% ±3.17 | 0.93 pts | Reported ±1 SE ranges overlap |
| Vibe Code Bench v1.1 | 81.33% ±3.05 | 76.24% ±3.82 | 5.09 pts | Reported ±1 SE ranges overlap |
| CyberBench v1.1 | 61.91% ±5.74 | 67.68% ±5.87 | 5.77 pts | Reported ±1 SE ranges overlap |
| Public Benefits Bench v1.1 | 66.03% ±1.23 | 66.85% ±1.23 | 0.81 pts | Reported ±1 SE ranges overlap |
Performance by category
| Category | Claude Sonnet 5 average | Grok 4.6 average |
|---|---|---|
| Legal | 43.58% | 50.07% |
| Finance | 65.63% | 64.50% |
| Healthcare | 61.80% | 65.62% |
| Math | 77.00% | 51.00% |
| Science | 2.86% | 11.43% |
| Academic | 88.22% | 92.05% |
| Education | 48.92% | 28.90% |
| Coding | 56.15% | 60.39% |
| Beta | 61.91% | 67.68% |
| Social Mobility | 66.03% | 66.85% |
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 | Grok 4.6 cost | Claude Sonnet 5 latency | Grok 4.6 latency |
|---|---|---|---|---|
| Harvey's Legal Agent Benchmark | $8.95 | $4.01 | 38m52s | 45m05s |
| Legal Research Bench | $2.72 | $1.53 | 25m46s | 25m04s |
| LegalBench | N/A | N/A | 4.90s | 27.42s |
| EMB | $10.29 | $3.06 | 49m46s | 52m59s |
| Finance Agent (v2) | $0.75 | $1.66 | 13m12s | 16m41s |
| MortgageTax | N/A | N/A | 28.67s | 8.55s |
| Tax Agent Bench | $1.79 | $0.98 | 19m19s | 10m58s |
| TaxEval v2 | N/A | N/A | 3m22s | 47.57s |
| MedCode | N/A | N/A | 2m15s | 2m07s |
| MedScribe | N/A | N/A | 4m12s | 70.94s |
| ProofBench v1.1 | $1.37 | $0.76 | 14m58s | 14m22s |
| Terminal-Bench Science | $18.70 | $4.82 | 2h55m | 1h18m |
| GPQA Diamond | N/A | N/A | 63.23s | 3m29s |
| MMLU Pro | N/A | N/A | 25.19s | 54.98s |
| SAGE | N/A | N/A | 7m14s | 3m22s |
| Code Migration | $35.31 | $14.58 | 1h57m | 1h30m |
| IOI | $12.89 | $8.18 | 59m44s | 3h57m |
| LiveCodeBench | N/A | N/A | 77.02s | 118.40s |
| SkillsBench | $2.90 | $0.92 | 15m12s | 13m28s |
| SWE-bench | $1.49 | $0.78 | 16m02s | 10m04s |
| Vibe Code Bench 1-100 | $71.15 | $26.27 | 4h16m | 2h11m |
| Vibe Code Bench v1.1 | $25.39 | $4.88 | 1h07m | 25m30s |
| CyberBench v1.1 | $1.89 | $4.07 | 16m57s | 36m52s |
| Public Benefits Bench v1.1 | $1.29 | $0.92 | 21m09s | 25m53s |
Results available only for Claude Sonnet 5
- Vals Index
- MMMU Pro
- ProgramBench
- Terminal-Bench 4.0
Results available only for Grok 4.6
- Vals RSI Index
- BioMysteryBench
- MysteryMechanism
- Time Horizon Index: KSP