Claude Sonnet 5.5 vs GPT-5.6 Sol: Benchmark Comparison
Claude Sonnet 5.5 has the higher score on 16 of 21 shared benchmarks; GPT-5.6 Sol leads on 4.
The largest observed score gap is 26.26 pts on Terminal-Bench 4.0 , where Claude Sonnet 5.5 leads.
Reported ±1 standard-error ranges overlap on 7 of 21 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-5.6 Sol | Gap | Reported uncertainty |
|---|---|---|---|---|
| Vals Index | 67.04% ±0.92 | 58.01% ±1.02 | 9.03 pts | Reported ±1 SE ranges do not overlap |
| Harvey's Legal Agent Benchmark | 2.92% ±1.36 | 2.50% ±0.83 | 0.42 pts | Reported ±1 SE ranges overlap |
| Legal Research Bench | 48.08% ±3.47 | 48.08% ±3.47 | 0.00 pts | Reported ±1 SE ranges overlap |
| EMB | 75.71% ±2.44 | 72.34% ±2.37 | 3.37 pts | Reported ±1 SE ranges overlap |
| Finance Agent (v2) | 58.10% ±0.67 | 53.76% ±0.85 | 4.35 pts | Reported ±1 SE ranges do not overlap |
| Tax Agent Bench | 41.96% ±3.56 | 31.09% ±3.34 | 10.87 pts | Reported ±1 SE ranges do not overlap |
| MedCode | 52.92% ±2.12 | 43.97% ±2.26 | 8.95 pts | Reported ±1 SE ranges do not overlap |
| MedScribe | 91.10% ±1.96 | 85.23% ±1.97 | 5.87 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 | 71.11% ±0.64 | 10.00 pts | Reported ±1 SE ranges do not overlap |
| MysteryMechanism | 49.10% ±3.36 | 33.33% ±3.17 | 15.77 pts | Reported ±1 SE ranges do not overlap |
| Terminal-Bench Science | 45.71% ±6.00 | 20.00% ±4.82 | 25.71 pts | Reported ±1 SE ranges do not overlap |
| SAGE | 51.77% ±3.41 | 52.56% ±3.42 | 0.79 pts | Reported ±1 SE ranges overlap |
| Code Migration | 69.83% ±4.26 | 52.92% ±4.35 | 16.91 pts | Reported ±1 SE ranges do not overlap |
| IOI | 83.06% ±3.74 | 91.17% ±4.51 | 8.11 pts | Reported ±1 SE ranges overlap |
| ProgramBench | 6.50% ±1.75 | 1.50% ±0.86 | 5.00 pts | Reported ±1 SE ranges do not overlap |
| Terminal-Bench 4.0 | 64.14% ±1.01 | 37.88% ±0.00 | 26.26 pts | Reported ±1 SE ranges do not overlap |
| Vibe Code Bench v1.1 | 92.39% ±1.26 | 80.50% ±3.72 | 11.90 pts | Reported ±1 SE ranges do not overlap |
| CyberBench v1.1 | 59.58% ±5.21 | 76.31% ±5.18 | 16.73 pts | Reported ±1 SE ranges do not overlap |
| SRE Bench | 30.15% ±2.84 | 30.53% ±2.85 | 0.38 pts | Reported ±1 SE ranges overlap |
| Public Benefits Bench v1.1 | 67.19% ±1.22 | 66.51% ±1.23 | 0.68 pts | Reported ±1 SE ranges overlap |
Performance by category
| Category | Claude Sonnet 5.5 average | GPT-5.6 Sol average |
|---|---|---|
| Legal | 25.50% | 25.29% |
| Finance | 58.59% | 52.39% |
| Healthcare | 72.01% | 64.60% |
| Math | 100.00% | 83.00% |
| Science | 58.64% | 41.48% |
| Education | 51.77% | 52.56% |
| Coding | 63.18% | 52.79% |
| Cyber | 44.87% | 53.42% |
| Social Mobility | 67.19% | 66.51% |
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-5.6 Sol cost | Claude Sonnet 5.5 latency | GPT-5.6 Sol latency |
|---|---|---|---|---|
| Vals Index | $21.34 | $14.24 | 1h18m | 37m23s |
| Harvey's Legal Agent Benchmark | $16.61 | $10.37 | 57m44s | 22m50s |
| Legal Research Bench | $14.83 | $21.61 | 1h33m | 1h17m |
| EMB | $8.79 | $6.01 | 36m45s | 9m50s |
| Finance Agent (v2) | $6.55 | $1.25 | 31m43s | 19m25s |
| Tax Agent Bench | $9.25 | $7.51 | 59m21s | 44m41s |
| MedCode | N/A | N/A | 3m59s | 96.59s |
| MedScribe | N/A | N/A | 5m07s | 94.40s |
| ProofBench v1.1 | $0.56 | $1.44 | 5m48s | 8m40s |
| BioMysteryBench | $3.74 | $1.93 | 18m06s | 18m18s |
| MysteryMechanism | $3.45 | $0.96 | 23m30s | 8m43s |
| Terminal-Bench Science | $30.43 | $7.32 | 2h39m | 1h44m |
| SAGE | N/A | N/A | 113.42s | 75.39s |
| Code Migration | $75.83 | $24.54 | 3h20m | 59m12s |
| IOI | $7.52 | $7.98 | 45m51s | 1h02m |
| ProgramBench | $35.86 | $15.29 | 2h07m | 29m24s |
| Terminal-Bench 4.0 | $16.51 | $7.98 | 1h22m | 33m46s |
| Vibe Code Bench v1.1 | $31.25 | $33.40 | 1h01m | 33m53s |
| CyberBench v1.1 | $4.42 | $3.05 | 33m53s | 10m47s |
| SRE Bench | $26.52 | $42.52 | 1h55m | 1h28m |
| Public Benefits Bench v1.1 | $4.25 | $8.85 | 1h15m | 1h01m |
Results available only for Claude Sonnet 5.5
None.
Results available only for GPT-5.6 Sol
- Vals RSI Index
- LegalBench
- MortgageTax
- TaxEval v2
- GPQA Diamond
- MMLU Pro
- MMMU Pro
- LiveCodeBench
- SkillsBench
- SWE-bench
- Vibe Code Bench 1-100
- CUA-bench
- Time Horizon Index: KSP