Model comparison

Claude Sonnet 5.5 vs Gemini 3.8 Flash: Benchmark Comparison

Claude Sonnet 5.5 has the higher score on 17 of 19 shared benchmarks; Gemini 3.8 Flash leads on 2.

The largest observed score gap is 52.00 pts on ProofBench v1.1 , where Claude Sonnet 5.5 leads.

Reported ±1 standard-error ranges overlap on 2 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 Gemini 3.8 Flash Gap Reported uncertainty
Vals Index 69.22% ±0.96 62.25% ±1.01 6.97 pts Reported ±1 SE ranges do not overlap
Harvey's Legal Agent Benchmark 2.92% ±1.36 10.00% ±2.42 7.08 pts Reported ±1 SE ranges do not overlap
Legal Research Bench 48.08% ±3.47 38.94% ±3.39 9.13 pts Reported ±1 SE ranges do not overlap
EMB 75.71% ±2.44 72.20% ±2.42 3.52 pts Reported ±1 SE ranges overlap
Finance Agent (v2) 58.10% ±0.67 61.44% ±0.13 3.33 pts Reported ±1 SE ranges do not overlap
Tax Agent Bench 73.39% ±2.98 66.77% ±3.15 6.62 pts Reported ±1 SE ranges do not overlap
MedCode 52.92% ±2.12 48.13% ±2.18 4.79 pts Reported ±1 SE ranges do not overlap
MedScribe 91.10% ±1.96 84.50% ±1.94 6.61 pts Reported ±1 SE ranges do not overlap
ProofBench v1.1 100.00% ±0.00 48.00% ±5.02 52.00 pts Reported ±1 SE ranges do not overlap
BioMysteryBench 81.11% ±0.64 62.22% ±1.70 18.89 pts Reported ±1 SE ranges do not overlap
MysteryMechanism 49.10% ±3.36 36.49% ±3.24 12.61 pts Reported ±1 SE ranges do not overlap
Terminal-Bench Science 38.57% ±5.86 8.57% ±3.37 30.00 pts Reported ±1 SE ranges do not overlap
SAGE 51.77% ±3.41 35.06% ±3.36 16.71 pts Reported ±1 SE ranges do not overlap
Code Migration 69.83% ±4.26 36.55% ±4.18 33.28 pts Reported ±1 SE ranges do not overlap
IOI 83.06% ±3.74 56.94% ±2.26 26.11 pts Reported ±1 SE ranges do not overlap
Terminal-Bench 4.0 53.03% ±1.51 13.13% ±2.52 39.90 pts Reported ±1 SE ranges do not overlap
Vibe Code Bench v1.1 92.39% ±1.26 78.65% ±3.88 13.74 pts Reported ±1 SE ranges do not overlap
CyberBench v1.1 59.58% ±5.21 43.75% ±2.21 15.83 pts Reported ±1 SE ranges do not overlap
Public Benefits Bench v1.1 67.19% ±1.22 65.29% ±1.24 1.89 pts Reported ±1 SE ranges overlap

Performance by category

Category Claude Sonnet 5.5 average Gemini 3.8 Flash average
Index 69.22% 62.25%
Legal 25.50% 24.47%
Finance 69.07% 66.80%
Healthcare 72.01% 66.32%
Math 100.00% 48.00%
Science 56.26% 35.76%
Education 51.77% 35.06%
Coding 74.58% 46.32%
Beta 59.58% 43.75%
Social Mobility 67.19% 65.29%

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 Gemini 3.8 Flash cost Claude Sonnet 5.5 latency Gemini 3.8 Flash latency
Vals Index $20.80 $5.39 1h10m 44m53s
Harvey's Legal Agent Benchmark $16.61 $3.66 57m44s 29m21s
Legal Research Bench $14.83 $1.63 1h33m 5m24s
EMB $8.79 $8.24 36m45s 12m49s
Finance Agent (v2) $6.55 $2.00 31m43s 3m21s
Tax Agent Bench $9.25 $0.86 59m21s 2m58s
MedCode N/A N/A 3m59s 42.89s
MedScribe N/A N/A 5m07s 21.14s
ProofBench v1.1 $0.56 $0.60 5m48s 6m09s
BioMysteryBench $3.74 $1.55 18m06s 5m27s
MysteryMechanism $3.45 $1.73 23m30s 5m22s
Terminal-Bench Science $31.98 $9.26 3h22m 48m18s
SAGE N/A N/A 113.42s 28.64s
Code Migration $75.83 $18.49 3h20m 2h37m
IOI $7.52 $3.98 45m51s 14m20s
Terminal-Bench 4.0 $19.33 $11.54 2h06m 2h22m
Vibe Code Bench v1.1 $31.25 $6.87 1h01m 8m39s
CyberBench v1.1 $4.42 $1.31 33m53s 9m49s
Public Benefits Bench v1.1 $4.25 $1.01 1h18m 8m37s

Results available only for Claude Sonnet 5.5

  • SRE Bench

Results available only for Gemini 3.8 Flash

  • Vals RSI Index
  • LegalBench
  • MortgageTax
  • TaxEval v2
  • GPQA Diamond
  • MMLU Pro
  • MMMU Pro
  • LiveCodeBench
  • ProgramBench
  • SkillsBench
  • SWE-bench
  • Vibe Code Bench 1-100
  • CUA-bench
  • Time Horizon Index: KSP
Model details Claude Sonnet 5.5 Model details Gemini 3.8 Flash