Claude Opus 4.8 vs GPT-5.6 Sol: Benchmark Comparison
Claude Opus 4.8 has the higher score on 13 of 26 shared benchmarks; GPT-5.6 Sol leads on 13.
The largest observed score gap is 15.71 pts on Terminal-Bench Science , where GPT-5.6 Sol leads.
Reported ±1 standard-error ranges overlap on 14 of 25 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 Opus 4.8 | GPT-5.6 Sol | Gap | Reported uncertainty |
|---|---|---|---|---|
| Vals Index | 55.10% ±1.01 | 58.01% ±1.02 | 2.91 pts | Reported ±1 SE ranges do not overlap |
| Vals RSI Index | 23.98% | 23.88% | 0.10 pts | Uncertainty comparison unavailable |
| Harvey's Legal Agent Benchmark | 9.58% ±2.15 | 2.50% ±0.83 | 7.08 pts | Reported ±1 SE ranges do not overlap |
| Legal Research Bench | 43.75% ±3.45 | 48.08% ±3.47 | 4.33 pts | Reported ±1 SE ranges overlap |
| LegalBench | 83.57% ±0.47 | 86.97% ±0.41 | 3.40 pts | Reported ±1 SE ranges do not overlap |
| EMB | 69.37% ±2.61 | 72.34% ±2.37 | 2.97 pts | Reported ±1 SE ranges overlap |
| Finance Agent (v2) | 53.92% ±0.16 | 53.76% ±0.85 | 0.16 pts | Reported ±1 SE ranges overlap |
| MortgageTax | 69.91% ±0.89 | 67.29% ±0.92 | 2.62 pts | Reported ±1 SE ranges do not overlap |
| Tax Agent Bench | 64.91% ±3.17 | 67.95% ±3.12 | 3.05 pts | Reported ±1 SE ranges overlap |
| TaxEval v2 | 75.63% ±0.84 | 74.78% ±0.86 | 0.86 pts | Reported ±1 SE ranges overlap |
| MedCode | 53.22% ±2.17 | 43.97% ±2.26 | 9.24 pts | Reported ±1 SE ranges do not overlap |
| MedScribe | 85.75% ±1.93 | 85.23% ±1.97 | 0.52 pts | Reported ±1 SE ranges overlap |
| Terminal-Bench Science | 4.29% ±2.44 | 20.00% ±4.82 | 15.71 pts | Reported ±1 SE ranges do not overlap |
| GPQA Diamond | 92.42% ±1.98 | 95.20% ±1.07 | 2.78 pts | Reported ±1 SE ranges overlap |
| MMLU Pro | 89.58% ±0.30 | 89.10% ±0.31 | 0.48 pts | Reported ±1 SE ranges overlap |
| MMMU Pro | 86.59% ±0.82 | 88.84% ±0.76 | 2.25 pts | Reported ±1 SE ranges do not overlap |
| SAGE | 54.79% ±3.35 | 52.56% ±3.42 | 2.22 pts | Reported ±1 SE ranges overlap |
| Code Migration | 47.25% ±4.18 | 52.92% ±4.35 | 5.67 pts | Reported ±1 SE ranges overlap |
| LiveCodeBench | 87.82% ±0.95 | 82.60% ±1.09 | 5.22 pts | Reported ±1 SE ranges do not overlap |
| ProgramBench | 1.00% ±0.70 | 1.50% ±0.86 | 0.50 pts | Reported ±1 SE ranges overlap |
| SkillsBench | 59.23% ±4.50 | 54.10% ±4.73 | 5.13 pts | Reported ±1 SE ranges overlap |
| SWE-bench | 88.60% ±1.42 | 96.20% ±0.86 | 7.60 pts | Reported ±1 SE ranges do not overlap |
| Terminal-Bench 4.0 | 23.23% ±1.34 | 37.88% ±0.00 | 14.65 pts | Reported ±1 SE ranges do not overlap |
| Vibe Code Bench v1.1 | 82.72% ±3.08 | 80.50% ±3.72 | 2.23 pts | Reported ±1 SE ranges overlap |
| Public Benefits Bench v1.1 | 68.13% ±1.21 | 66.51% ±1.23 | 1.62 pts | Reported ±1 SE ranges overlap |
| Time Horizon Index: KSP | 11.83% ±0.00 | 23.83% ±0.00 | 12.00 pts | Reported ±1 SE ranges do not overlap |
Performance by category
| Category | Claude Opus 4.8 average | GPT-5.6 Sol average |
|---|---|---|
| Legal | 45.63% | 45.85% |
| Finance | 66.75% | 67.22% |
| Healthcare | 69.49% | 64.60% |
| Science | 4.29% | 20.00% |
| Academic | 89.53% | 91.05% |
| Education | 54.79% | 52.56% |
| Coding | 55.69% | 57.96% |
| Social Mobility | 68.13% | 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 Opus 4.8 cost | GPT-5.6 Sol cost | Claude Opus 4.8 latency | GPT-5.6 Sol latency |
|---|---|---|---|---|
| Vals Index | $13.14 | $14.24 | 39m48s | 37m23s |
| Vals RSI Index | $465.25 | $347.10 | 90h00m | 90h00m |
| Harvey's Legal Agent Benchmark | $10.22 | $10.37 | 24m08s | 22m50s |
| Legal Research Bench | $2.82 | $21.61 | 11m54s | 1h17m |
| LegalBench | N/A | N/A | 3.11s | 6.20s |
| EMB | $12.06 | $6.01 | 43m52s | 9m50s |
| Finance Agent (v2) | $4.22 | $1.25 | 9m07s | 19m25s |
| MortgageTax | N/A | N/A | 20.81s | 12.62s |
| Tax Agent Bench | $1.69 | $7.51 | 7m59s | 44m41s |
| TaxEval v2 | N/A | N/A | 70.33s | 62.03s |
| MedCode | N/A | N/A | 105.92s | 96.59s |
| MedScribe | N/A | N/A | 82.36s | 94.40s |
| Terminal-Bench Science | $23.14 | $7.32 | 1h55m | 1h44m |
| GPQA Diamond | N/A | N/A | 2m14s | 58.30s |
| MMLU Pro | N/A | N/A | 23.44s | 15.65s |
| MMMU Pro | N/A | N/A | 49.04s | 26.20s |
| SAGE | N/A | N/A | 3m06s | 75.39s |
| Code Migration | $30.51 | $24.54 | 1h15m | 59m12s |
| LiveCodeBench | N/A | N/A | 2m40s | 56.51s |
| ProgramBench | $31.27 | $15.29 | 1h32m | 29m24s |
| SkillsBench | $4.77 | $4.14 | 15m30s | 13m07s |
| SWE-bench | $1.92 | $1.15 | 9m27s | 3m02s |
| Terminal-Bench 4.0 | $17.14 | $7.98 | 1h11m | 33m46s |
| Vibe Code Bench v1.1 | $26.88 | $33.40 | 1h16m | 33m53s |
| Public Benefits Bench v1.1 | $1.89 | $8.85 | 12m48s | 1h01m |
| Time Horizon Index: KSP | $1499.32 | $2027.20 | N/A | N/A |
Results available only for Claude Opus 4.8
None.
Results available only for GPT-5.6 Sol
- ProofBench v1.1
- BioMysteryBench
- MysteryMechanism
- IOI
- Vibe Code Bench 1-100
- CyberBench v1.1
- SRE Bench
- CUA-bench