Wonder vs GPT-6 Sol: Benchmark Comparison
Wonder has the higher score on 6 of 12 shared benchmarks; GPT-6 Sol leads on 6.
The largest observed score gap is 21.64 pts on Legal Research Bench , where Wonder leads.
Reported ±1 standard-error ranges overlap on 6 of 12 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-6 Sol | Gap | Reported uncertainty |
|---|---|---|---|---|
| Vals Science Index | 39.12% ±1.96 | 45.00% ±2.26 | 5.87 pts | Reported ±1 SE ranges do not overlap |
| Legal Research Bench | 50.48% ±3.48 | 28.85% ±3.15 | 21.64 pts | Reported ±1 SE ranges do not overlap |
| Tax Agent Bench | 36.27% ±3.47 | 15.00% ±2.58 | 21.27 pts | Reported ±1 SE ranges do not 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 | 74.81% ±2.43 | 7.78 pts | Reported ±1 SE ranges do not overlap |
| MysteryMechanism | 27.48% ±3.00 | 30.18% ±3.09 | 2.70 pts | Reported ±1 SE ranges overlap |
| Terminal-Bench-Science | 22.86% ±5.05 | 30.00% ±5.52 | 7.14 pts | Reported ±1 SE ranges overlap |
| IOI | 66.94% ±4.91 | 82.61% ±9.24 | 15.67 pts | Reported ±1 SE ranges do not overlap |
| Terminal-Bench 4.0 | 49.49% ±2.20 | 44.44% ±3.54 | 5.05 pts | Reported ±1 SE ranges overlap |
| Vibe Code Bench v1.1 | 83.09% ±2.49 | 87.82% ±2.53 | 4.73 pts | Reported ±1 SE ranges overlap |
| CyberBench v1.1 | 80.36% ±5.13 | 77.98% ±5.11 | 2.38 pts | Reported ±1 SE ranges overlap |
| Public Benefits Bench v1.1 | 70.23% ±1.19 | 56.63% ±1.29 | 13.60 pts | Reported ±1 SE ranges do not overlap |
Performance by category
| Category | Wonder average | GPT-6 Sol average |
|---|---|---|
| Legal | 50.48% | 28.85% |
| Finance | 36.27% | 15.00% |
| Math | 84.00% | 83.00% |
| Science | 39.12% | 45.00% |
| Coding | 66.51% | 71.63% |
| Cyber | 80.36% | 77.98% |
| Social Mobility | 70.23% | 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 | Wonder cost | GPT-6 Sol cost | Wonder latency | GPT-6 Sol latency |
|---|---|---|---|---|
| Vals Science Index | $8.74 | $2.31 | 43m08s | 31m16s |
| Legal Research Bench | $3.98 | $4.90 | 15m27s | 24m09s |
| Tax Agent Bench | $1.88 | $2.28 | 8m13s | 17m32s |
| ProofBench v1.1 | $1.31 | $0.48 | 15m29s | 5m05s |
| BioMysteryBench | $3.23 | $0.66 | 18m02s | 5m24s |
| MysteryMechanism | $3.39 | $0.46 | 18m17s | 6m13s |
| Terminal-Bench-Science | $19.61 | $5.82 | 1h33m | 1h22m |
| IOI | $4.54 | $2.70 | 25m02s | 30m38s |
| Terminal-Bench 4.0 | $14.66 | $5.79 | 54m41s | 34m56s |
| Vibe Code Bench v1.1 | $1.80 | $26.36 | 5m50s | 38m56s |
| CyberBench v1.1 | $3.66 | $1.34 | 13m45s | 12m14s |
| Public Benefits Bench v1.1 | $2.46 | $2.49 | 19m44s | 31m02s |
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
- Vibe Code Bench 1-100
- CUA-bench
Results available only for GPT-6 Sol
- Vals RSI Index
- MedCode
- MedScribe
- SAGE
- ProgramBench
- SRE Bench