Claude Opus 5 vs GPT-5.6 Terra: Benchmark Comparison
Claude Opus 5 has the higher score on 25 of 28 shared benchmarks; GPT-5.6 Terra leads on 3.
The largest observed score gap is 30.81 pts on Terminal-Bench 4.0 , where Claude Opus 5 leads.
Reported ±1 standard-error ranges overlap on 7 of 28 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 5 | GPT-5.6 Terra | Gap | Reported uncertainty |
|---|---|---|---|---|
| Vals Index | 63.67% ±0.96 | 53.09% ±1.29 | 10.59 pts | Reported ±1 SE ranges do not overlap |
| Harvey's Legal Agent Benchmark | 6.67% ±1.65 | 0.83% ±0.83 | 5.83 pts | Reported ±1 SE ranges do not overlap |
| Legal Research Bench | 55.29% ±3.46 | 41.35% ±3.42 | 13.94 pts | Reported ±1 SE ranges do not overlap |
| LegalBench | 86.97% ±0.42 | 85.11% ±0.45 | 1.87 pts | Reported ±1 SE ranges do not overlap |
| EMB | 73.56% ±2.24 | 66.20% ±3.11 | 7.35 pts | Reported ±1 SE ranges do not overlap |
| Finance Agent (v2) | 58.63% ±0.08 | 54.44% ±2.07 | 4.20 pts | Reported ±1 SE ranges do not overlap |
| MortgageTax | 72.06% ±0.88 | 67.33% ±0.93 | 4.73 pts | Reported ±1 SE ranges do not overlap |
| Tax Agent Bench | 75.06% ±2.96 | 65.20% ±3.19 | 9.86 pts | Reported ±1 SE ranges do not overlap |
| TaxEval v2 | 75.14% ±0.83 | 76.17% ±0.84 | 1.02 pts | Reported ±1 SE ranges overlap |
| MedCode | 63.57% ±1.99 | 43.41% ±2.17 | 20.16 pts | Reported ±1 SE ranges do not overlap |
| MedScribe | 90.98% ±1.92 | 82.87% ±1.95 | 8.12 pts | Reported ±1 SE ranges do not overlap |
| ProofBench v1.1 | 99.00% ±1.00 | 74.00% ±4.41 | 25.00 pts | Reported ±1 SE ranges do not overlap |
| Terminal-Bench Science | 27.14% ±5.35 | 10.00% ±3.61 | 17.14 pts | Reported ±1 SE ranges do not overlap |
| GPQA Diamond | 93.43% ±1.24 | 90.91% ±1.91 | 2.52 pts | Reported ±1 SE ranges overlap |
| MMLU Pro | 91.59% ±0.28 | 86.66% ±0.33 | 4.93 pts | Reported ±1 SE ranges do not overlap |
| MMMU Pro | 89.88% ±0.72 | 86.47% ±0.82 | 3.41 pts | Reported ±1 SE ranges do not overlap |
| SAGE | 49.43% ±3.29 | 47.00% ±3.40 | 2.42 pts | Reported ±1 SE ranges overlap |
| Code Migration | 57.47% ±4.37 | 47.80% ±4.28 | 9.67 pts | Reported ±1 SE ranges do not overlap |
| IOI | 84.33% ±9.96 | 87.61% ±6.53 | 3.28 pts | Reported ±1 SE ranges overlap |
| LiveCodeBench | 89.03% ±0.91 | 85.93% ±1.02 | 3.10 pts | Reported ±1 SE ranges do not overlap |
| ProgramBench | 3.00% ±1.21 | 0.50% ±0.50 | 2.50 pts | Reported ±1 SE ranges do not overlap |
| SkillsBench | 60.44% ±4.58 | 58.90% ±4.47 | 1.55 pts | Reported ±1 SE ranges overlap |
| SWE-bench | 97.00% ±0.76 | 95.40% ±0.94 | 1.60 pts | Reported ±1 SE ranges overlap |
| Terminal-Bench 4.0 | 53.53% ±1.34 | 22.73% ±2.31 | 30.81 pts | Reported ±1 SE ranges do not overlap |
| Vibe Code Bench 1-100 | 28.53% ±4.33 | 14.82% ±3.12 | 13.71 pts | Reported ±1 SE ranges do not overlap |
| Vibe Code Bench v1.1 | 88.40% ±3.00 | 74.59% ±4.20 | 13.81 pts | Reported ±1 SE ranges do not overlap |
| CyberBench v1.1 | 65.36% ±5.55 | 72.08% ±5.41 | 6.73 pts | Reported ±1 SE ranges overlap |
| Public Benefits Bench v1.1 | 76.93% ±1.10 | 62.38% ±1.26 | 14.55 pts | Reported ±1 SE ranges do not overlap |
Performance by category
| Category | Claude Opus 5 average | GPT-5.6 Terra average |
|---|---|---|
| Legal | 49.64% | 42.43% |
| Finance | 70.89% | 65.87% |
| Healthcare | 77.28% | 63.14% |
| Math | 99.00% | 74.00% |
| Science | 27.14% | 10.00% |
| Academic | 91.64% | 88.01% |
| Education | 49.43% | 47.00% |
| Coding | 62.42% | 54.25% |
| Cyber | 65.36% | 72.08% |
| Social Mobility | 76.93% | 62.38% |
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 5 cost | GPT-5.6 Terra cost | Claude Opus 5 latency | GPT-5.6 Terra latency |
|---|---|---|---|---|
| Vals Index | $19.31 | $5.78 | 57m41s | 35m39s |
| Harvey's Legal Agent Benchmark | $23.67 | $3.20 | 55m37s | 17m05s |
| Legal Research Bench | $6.76 | $7.91 | 31m01s | 1h10m |
| LegalBench | N/A | N/A | 4.42s | 2.78s |
| EMB | $6.07 | $2.19 | 25m59s | 13m33s |
| Finance Agent (v2) | $5.12 | $3.63 | 9m58s | 24m10s |
| MortgageTax | N/A | N/A | 7.98s | 6.32s |
| Tax Agent Bench | $5.14 | $4.83 | 16m48s | 1h01m |
| TaxEval v2 | N/A | N/A | 58.34s | 37.28s |
| MedCode | N/A | N/A | 24.50s | 18.41s |
| MedScribe | N/A | N/A | 76.56s | 35.53s |
| ProofBench v1.1 | $2.11 | $1.27 | 9m51s | 12m57s |
| Terminal-Bench Science | $32.54 | $5.19 | 2h12m | 2h04m |
| GPQA Diamond | N/A | N/A | 39.59s | 36.76s |
| MMLU Pro | N/A | N/A | 9.60s | 8.17s |
| MMMU Pro | N/A | N/A | 30.03s | 13.03s |
| SAGE | N/A | N/A | 56.17s | 40.06s |
| Code Migration | $60.51 | $8.13 | 2h51m | 48m34s |
| IOI | $16.48 | $8.66 | 56m42s | 1h23m |
| LiveCodeBench | N/A | N/A | 59.22s | 44.33s |
| ProgramBench | $60.29 | $5.38 | 2h55m | 33m36s |
| SkillsBench | $2.51 | $1.75 | 8m25s | 6m55s |
| SWE-bench | $1.29 | $0.40 | 9m37s | 3m00s |
| Terminal-Bench 4.0 | $18.60 | $5.60 | 1h07m | 31m29s |
| Vibe Code Bench 1-100 | $41.49 | $33.33 | 1h22m | 1h49m |
| Vibe Code Bench v1.1 | $33.88 | $7.89 | 1h28m | 19m12s |
| CyberBench v1.1 | $2.87 | $3.31 | 19m25s | 17m04s |
| Public Benefits Bench v1.1 | $3.95 | $1.20 | 30m09s | 19m31s |
Results available only for Claude Opus 5
- Vals RSI Index
- BioMysteryBench
- MysteryMechanism
- SRE Bench
- CUA-bench
- Time Horizon Index: KSP
Results available only for GPT-5.6 Terra
None.