Model comparison

Claude Sonnet 5.5 vs GPT-6 Sol: Benchmark Comparison

Claude Sonnet 5.5 has the higher score on 18 of 19 shared benchmarks; GPT-6 Sol leads on 1.

The largest observed score gap is 20.35 pts on Tax Agent Bench , where Claude Sonnet 5.5 leads.

Reported ±1 standard-error ranges overlap on 4 of 19 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 5.5 GPT-6 Sol Gap Reported uncertainty
Vals Index 69.22% ±0.96 62.57% ±0.95 6.65 pts Reported ±1 SE ranges do not overlap
Harvey's Legal Agent Benchmark 2.92% ±1.36 1.67% ±0.82 1.25 pts Reported ±1 SE ranges overlap
Legal Research Bench 48.08% ±3.47 28.85% ±3.15 19.23 pts Reported ±1 SE ranges do not overlap
EMB 75.71% ±2.44 71.53% ±2.35 4.18 pts Reported ±1 SE ranges overlap
Finance Agent (v2) 58.10% ±0.67 49.05% ±0.58 9.05 pts Reported ±1 SE ranges do not overlap
Tax Agent Bench 73.39% ±2.98 53.05% ±3.29 20.35 pts Reported ±1 SE ranges do not overlap
MedCode 52.92% ±2.12 47.07% ±2.12 5.85 pts Reported ±1 SE ranges do not overlap
MedScribe 91.10% ±1.96 82.03% ±1.94 9.07 pts Reported ±1 SE ranges do not overlap
ProofBench v1.1 100.00% ±0.00 83.00% ±3.77 17.00 pts Reported ±1 SE ranges do not overlap
BioMysteryBench 81.11% ±0.64 74.81% ±2.43 6.30 pts Reported ±1 SE ranges do not overlap
MysteryMechanism 49.10% ±3.36 30.18% ±3.09 18.92 pts Reported ±1 SE ranges do not overlap
Terminal-Bench Science 38.57% ±5.86 31.43% ±5.59 7.14 pts Reported ±1 SE ranges overlap
SAGE 51.77% ±3.41 44.79% ±3.36 6.98 pts Reported ±1 SE ranges do not overlap
Code Migration 69.83% ±4.26 57.20% ±4.21 12.63 pts Reported ±1 SE ranges do not overlap
IOI 83.06% ±3.74 82.61% ±9.24 0.44 pts Reported ±1 SE ranges overlap
Terminal-Bench 4.0 53.03% ±1.51 34.34% ±2.02 18.69 pts Reported ±1 SE ranges do not overlap
Vibe Code Bench v1.1 92.39% ±1.26 87.82% ±2.53 4.57 pts Reported ±1 SE ranges do not overlap
CyberBench v1.1 59.58% ±5.21 77.98% ±5.11 18.39 pts Reported ±1 SE ranges do not overlap
Public Benefits Bench v1.1 67.19% ±1.22 56.63% ±1.29 10.55 pts Reported ±1 SE ranges do not overlap

Performance by category

Category Claude Sonnet 5.5 average GPT-6 Sol average
Index 69.22% 62.57%
Legal 25.50% 15.26%
Finance 69.07% 57.88%
Healthcare 72.01% 64.55%
Math 100.00% 83.00%
Science 56.26% 45.47%
Education 51.77% 44.79%
Coding 74.58% 65.49%
Beta 59.58% 77.98%
Social Mobility 67.19% 56.63%

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 5.5 cost GPT-6 Sol cost Claude Sonnet 5.5 latency GPT-6 Sol latency
Vals Index $20.80 $7.56 1h10m 26m34s
Harvey's Legal Agent Benchmark $16.61 $3.36 57m44s 13m40s
Legal Research Bench $14.83 $4.90 1h33m 24m09s
EMB $8.79 $1.28 36m45s 10m22s
Finance Agent (v2) $6.55 $2.12 31m43s 9m00s
Tax Agent Bench $9.25 $2.28 59m21s 17m32s
MedCode N/A N/A 3m59s 62.71s
MedScribe N/A N/A 5m07s 81.14s
ProofBench v1.1 $0.56 $0.48 5m48s 5m05s
BioMysteryBench $3.74 $0.66 18m06s 5m24s
MysteryMechanism $3.45 $0.46 23m30s 6m13s
Terminal-Bench Science $31.98 $9.05 3h22m 1h21m
SAGE N/A N/A 113.42s 31.42s
Code Migration $75.83 $15.68 3h20m 1h26m
IOI $7.52 $2.70 45m51s 30m38s
Terminal-Bench 4.0 $19.33 $4.14 2h06m 43m10s
Vibe Code Bench v1.1 $31.25 $26.36 1h01m 38m56s
CyberBench v1.1 $4.42 $1.34 33m53s 12m14s
Public Benefits Bench v1.1 $4.25 $2.50 1h18m 33m34s

Results available only for Claude Sonnet 5.5

  • SRE Bench

Results available only for GPT-6 Sol

  • Vals RSI Index
  • ProgramBench
Model details Claude Sonnet 5.5 Model details GPT-6 Sol