Claude Fable 5.1 vs GPT-6 Sol: Benchmark Comparison

Claude Fable 5.1 has the higher score on 14 of 16 shared benchmarks; GPT-6 Sol leads on 2.

The largest observed score gap is 35.75 pts on Tax Agent Bench , where Claude Fable 5.1 leads.

Reported ±1 standard-error ranges overlap on 4 of 15 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 Fable 5.1 GPT-6 Sol Gap Reported uncertainty
Vals RSI Index 36.09% 28.11% 7.98 pts Uncertainty comparison unavailable
Legal Research Bench 55.29% ±3.46 28.85% ±3.15 26.44 pts Reported ±1 SE ranges do not overlap
Tax Agent Bench 50.75% ±3.61 15.00% ±2.58 35.75 pts Reported ±1 SE ranges do not overlap
MedCode 53.51% ±2.17 47.07% ±2.12 6.44 pts Reported ±1 SE ranges do not overlap
MedScribe 91.29% ±1.95 82.03% ±1.94 9.26 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
MysteryMechanism 47.75% ±3.36 30.18% ±3.09 17.57 pts Reported ±1 SE ranges do not overlap
Terminal-Bench-Science 40.00% ±5.90 30.00% ±5.52 10.00 pts Reported ±1 SE ranges overlap
IOI 90.78% ±4.65 82.61% ±9.24 8.17 pts Reported ±1 SE ranges overlap
ProgramBench 7.00% ±1.81 2.00% ±0.99 5.00 pts Reported ±1 SE ranges do not overlap
Terminal-Bench 4.0 58.08% ±3.31 44.44% ±3.54 13.64 pts Reported ±1 SE ranges do not overlap
Vibe Code Bench v1.1 90.26% ±1.57 87.82% ±2.53 2.44 pts Reported ±1 SE ranges overlap
CyberBench v1.1 70.42% ±5.43 77.98% ±5.11 7.56 pts Reported ±1 SE ranges overlap
SRE Bench 22.90% ±2.60 30.53% ±2.85 7.63 pts Reported ±1 SE ranges do not overlap
Financial Coaching Bench 75.71% ±1.54 62.61% ±1.72 13.10 pts Reported ±1 SE ranges do not overlap
Public Benefits Bench v1.1 74.90% ±1.13 56.63% ±1.29 18.27 pts Reported ±1 SE ranges do not overlap

Performance by category

Category Claude Fable 5.1 average GPT-6 Sol average
Legal 55.29% 28.85%
Finance 50.75% 15.00%
Healthcare 72.40% 64.55%
Math 100.00% 83.00%
Science 43.87% 30.09%
Coding 61.53% 54.22%
Cyber 46.66% 54.25%
Social Mobility 75.31% 59.62%

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 Fable 5.1 cost GPT-6 Sol cost Claude Fable 5.1 latency GPT-6 Sol latency
Vals RSI Index $298.22 $665.11 90h00m 90h00m
Legal Research Bench $23.06 $4.90 57m29s 24m09s
Tax Agent Bench $13.18 $2.28 28m34s 17m32s
MedCode N/A N/A 3m35s 62.71s
MedScribe N/A N/A 3m08s 81.14s
ProofBench v1.1 $2.70 $0.48 7m58s 5m05s
MysteryMechanism $5.63 $0.46 17m41s 6m13s
Terminal-Bench-Science $38.01 $5.82 2h36m 1h22m
IOI $11.20 $2.70 30m06s 30m38s
ProgramBench $58.52 $11.67 2h27m 46m11s
Terminal-Bench 4.0 $17.18 $5.79 1h00m 34m56s
Vibe Code Bench v1.1 $33.37 $26.36 57m40s 38m56s
CyberBench v1.1 $4.01 $1.34 18m32s 12m14s
SRE Bench $32.23 $20.65 2h03m 1h03m
Financial Coaching Bench $1.14 $0.23 5m48s 6m51s
Public Benefits Bench v1.1 $7.48 $2.49 48m02s 31m02s

Results available only for Claude Fable 5.1

  • Immigration
  • LegalBench
  • EMB v1.1
  • MortgageTax
  • TaxEval v2
  • Robo-bench
  • GPQA Diamond
  • MMLU Pro
  • MMMU Pro
  • LiveCodeBench
  • SkillsBench
  • SWE-bench Multimodal
  • CUA-bench
  • Time Horizon Index: KSP

Results available only for GPT-6 Sol

  • Vals Science Index
  • BioMysteryBench
Model details Claude Fable 5.1 Model details GPT-6 Sol