Wonder vs GPT-5.6 Sol: Benchmark Comparison
Wonder has the higher score on 9 of 14 shared benchmarks; GPT-5.6 Sol leads on 5.
The largest observed score gap is 24.22 pts on IOI , where GPT-5.6 Sol leads.
Reported ±1 standard-error ranges overlap on 9 of 13 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 | Wonder | GPT-5.6 Sol | Gap | Reported uncertainty |
|---|---|---|---|---|
| Vals Science Index | 39.12% ±1.96 | 41.48% ±1.93 | 2.36 pts | Reported ±1 SE ranges overlap |
| Legal Research Bench | 50.48% ±3.48 | 48.08% ±3.47 | 2.40 pts | Reported ±1 SE ranges overlap |
| Tax Agent Bench | 36.27% ±3.47 | 31.09% ±3.34 | 5.18 pts | Reported ±1 SE ranges overlap |
| ProofBench v1.1 | 84.00% ±3.69 | 83.00% ±3.77 | 1.00 pts | Reported ±1 SE ranges overlap |
| BioMysteryBench | 67.04% ±0.37 | 71.11% ±0.64 | 4.07 pts | Reported ±1 SE ranges do not overlap |
| MysteryMechanism | 27.48% ±3.00 | 33.33% ±3.17 | 5.86 pts | Reported ±1 SE ranges overlap |
| Terminal-Bench-Science | 22.86% ±5.05 | 20.00% ±4.82 | 2.86 pts | Reported ±1 SE ranges overlap |
| IOI | 66.94% ±4.91 | 91.17% ±4.51 | 24.22 pts | Reported ±1 SE ranges do not overlap |
| Terminal-Bench 4.0 | 49.49% ±2.20 | 37.88% ±0.00 | 11.62 pts | Reported ±1 SE ranges do not overlap |
| Vibe Code Bench 1-100 | 26.55% ±4.33 | 20.01% ±3.63 | 6.54 pts | Reported ±1 SE ranges overlap |
| Vibe Code Bench v1.1 | 83.09% ±2.49 | 80.50% ±3.72 | 2.59 pts | Reported ±1 SE ranges overlap |
| CyberBench v1.1 | 80.36% ±5.13 | 76.31% ±5.18 | 4.05 pts | Reported ±1 SE ranges overlap |
| Public Benefits Bench v1.1 | 70.23% ±1.19 | 66.51% ±1.23 | 3.72 pts | Reported ±1 SE ranges do not overlap |
| CUA-bench | 7.17% | 8.33% | 1.17 pts | Uncertainty comparison unavailable |
Performance by category
| Category | Wonder average | GPT-5.6 Sol average |
|---|---|---|
| Legal | 50.48% | 48.08% |
| Finance | 36.27% | 31.09% |
| Math | 84.00% | 83.00% |
| Science | 39.12% | 41.48% |
| Coding | 56.52% | 57.39% |
| Cyber | 80.36% | 76.31% |
| Social Mobility | 70.23% | 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 | Wonder cost | GPT-5.6 Sol cost | Wonder latency | GPT-5.6 Sol latency |
|---|---|---|---|---|
| Vals Science Index | $8.74 | $3.40 | 43m08s | 43m32s |
| Legal Research Bench | $3.98 | $21.61 | 15m27s | 1h17m |
| Tax Agent Bench | $1.88 | $7.51 | 8m13s | 44m41s |
| ProofBench v1.1 | $1.31 | $1.44 | 15m29s | 8m40s |
| BioMysteryBench | $3.23 | $1.93 | 18m02s | 18m18s |
| MysteryMechanism | $3.39 | $0.96 | 18m17s | 8m43s |
| Terminal-Bench-Science | $19.61 | $7.32 | 1h33m | 1h44m |
| IOI | $4.54 | $7.98 | 25m02s | 1h02m |
| Terminal-Bench 4.0 | $14.66 | $7.98 | 54m41s | 33m46s |
| Vibe Code Bench 1-100 | $13.72 | $47.43 | 31m08s | 2h37m |
| Vibe Code Bench v1.1 | $1.80 | $33.40 | 5m50s | 33m53s |
| CyberBench v1.1 | $3.66 | $3.05 | 13m45s | 10m47s |
| Public Benefits Bench v1.1 | $2.46 | $8.85 | 19m44s | 1h01m |
| CUA-bench | $597.21 | $206.99 | 30.56s | 18.80s |
Results available only for Wonder
- Vals Finance Index
- Vals Index
- Vals Legal Index
- HLAB-Vals-1.1
- EMB v1.1
- Finance Agent v2.1
- Code Migration v1.1
- SWE-bench Multimodal
Results available only for GPT-5.6 Sol
- Vals RSI Index
- LegalBench
- MortgageTax
- TaxEval v2
- MedCode
- MedScribe
- GPQA Diamond
- MMLU Pro
- MMMU Pro
- SAGE
- LiveCodeBench
- ProgramBench
- SkillsBench
- SWE-bench
- SRE Bench
- Time Horizon Index: KSP