Claude Opus 4.8 vs Gemini 3.7 Flash: Benchmark Comparison
Claude Opus 4.8 has the higher score on 14 of 24 shared benchmarks; Gemini 3.7 Flash leads on 10.
The largest observed score gap is 12.45 pts on Code Migration , where Claude Opus 4.8 leads.
Reported ±1 standard-error ranges overlap on 11 of 23 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 | Gemini 3.7 Flash | Gap | Reported uncertainty |
|---|---|---|---|---|
| Vals Index | 55.10% ±1.01 | 51.27% ±1.10 | 3.83 pts | Reported ±1 SE ranges do not overlap |
| Vals RSI Index | 23.98% | 18.27% | 5.71 pts | Uncertainty comparison unavailable |
| Harvey's Legal Agent Benchmark | 9.58% ±2.15 | 8.75% ±2.15 | 0.83 pts | Reported ±1 SE ranges overlap |
| Legal Research Bench | 43.75% ±3.45 | 34.62% ±3.31 | 9.13 pts | Reported ±1 SE ranges do not overlap |
| LegalBench | 83.57% ±0.47 | 87.26% ±0.42 | 3.69 pts | Reported ±1 SE ranges do not overlap |
| EMB | 69.37% ±2.61 | 71.33% ±2.26 | 1.95 pts | Reported ±1 SE ranges overlap |
| Finance Agent (v2) | 53.92% ±0.16 | 59.04% ±0.27 | 5.12 pts | Reported ±1 SE ranges do not overlap |
| MortgageTax | 69.91% ±0.89 | 66.65% ±0.92 | 3.26 pts | Reported ±1 SE ranges do not overlap |
| Tax Agent Bench | 64.91% ±3.17 | 57.66% ±3.31 | 7.24 pts | Reported ±1 SE ranges do not overlap |
| TaxEval v2 | 75.63% ±0.84 | 74.73% ±0.85 | 0.90 pts | Reported ±1 SE ranges overlap |
| MedCode | 53.22% ±2.17 | 53.39% ±2.12 | 0.17 pts | Reported ±1 SE ranges overlap |
| MedScribe | 85.75% ±1.93 | 83.94% ±2.00 | 1.81 pts | Reported ±1 SE ranges overlap |
| Terminal-Bench Science | 4.29% ±2.44 | 7.14% ±3.10 | 2.86 pts | Reported ±1 SE ranges overlap |
| GPQA Diamond | 92.42% ±1.98 | 93.94% ±1.49 | 1.52 pts | Reported ±1 SE ranges overlap |
| MMLU Pro | 89.58% ±0.30 | 90.12% ±0.30 | 0.54 pts | Reported ±1 SE ranges overlap |
| MMMU Pro | 86.59% ±0.82 | 88.96% ±0.75 | 2.37 pts | Reported ±1 SE ranges do not overlap |
| SAGE | 54.79% ±3.35 | 49.23% ±3.38 | 5.56 pts | Reported ±1 SE ranges overlap |
| Code Migration | 47.25% ±4.18 | 34.80% ±4.22 | 12.45 pts | Reported ±1 SE ranges do not overlap |
| LiveCodeBench | 87.82% ±0.95 | 88.65% ±0.92 | 0.83 pts | Reported ±1 SE ranges overlap |
| ProgramBench | 1.00% ±0.70 | 0.00% ±0.00 | 1.00 pts | Reported ±1 SE ranges do not overlap |
| SkillsBench | 59.23% ±4.50 | 65.89% ±4.55 | 6.66 pts | Reported ±1 SE ranges overlap |
| SWE-bench | 88.60% ±1.42 | 80.80% ±1.76 | 7.80 pts | Reported ±1 SE ranges do not overlap |
| Terminal-Bench 4.0 | 23.23% ±1.34 | 12.12% ±3.03 | 11.11 pts | Reported ±1 SE ranges do not overlap |
| Vibe Code Bench v1.1 | 82.72% ±3.08 | 70.39% ±4.84 | 12.33 pts | Reported ±1 SE ranges do not overlap |
Performance by category
| Category | Claude Opus 4.8 average | Gemini 3.7 Flash average |
|---|---|---|
| Legal | 45.63% | 43.54% |
| Finance | 66.75% | 65.88% |
| Healthcare | 69.49% | 68.67% |
| Science | 4.29% | 7.14% |
| Academic | 89.53% | 91.01% |
| Education | 54.79% | 49.23% |
| Coding | 55.69% | 50.38% |
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 4.8 cost | Gemini 3.7 Flash cost | Claude Opus 4.8 latency | Gemini 3.7 Flash latency |
|---|---|---|---|---|
| Vals Index | $13.14 | $4.71 | 39m48s | 47m27s |
| Vals RSI Index | $465.25 | $180.16 | 90h00m | 90h00m |
| Harvey's Legal Agent Benchmark | $10.22 | $2.56 | 24m08s | 7m32s |
| Legal Research Bench | $2.82 | $0.76 | 11m54s | 4m49s |
| LegalBench | N/A | N/A | 3.11s | 2.23s |
| EMB | $12.06 | $5.69 | 43m52s | 10m51s |
| Finance Agent (v2) | $4.22 | $1.48 | 9m07s | 3m01s |
| MortgageTax | N/A | N/A | 20.81s | 4.11s |
| Tax Agent Bench | $1.69 | $0.48 | 7m59s | 87.24s |
| TaxEval v2 | N/A | N/A | 70.33s | 7.83s |
| MedCode | N/A | N/A | 105.92s | 9.16s |
| MedScribe | N/A | N/A | 82.36s | 16.82s |
| Terminal-Bench Science | $23.14 | $6.26 | 1h55m | 1h18m |
| GPQA Diamond | N/A | N/A | 2m14s | 8.77s |
| MMLU Pro | N/A | N/A | 23.44s | 4.16s |
| MMMU Pro | N/A | N/A | 49.04s | 6.28s |
| SAGE | N/A | N/A | 3m06s | 9.79s |
| Code Migration | $30.51 | $21.46 | 1h15m | 2h39m |
| LiveCodeBench | N/A | N/A | 2m40s | 14.53s |
| ProgramBench | $31.27 | $7.13 | 1h32m | 21m30s |
| SkillsBench | $4.77 | $1.80 | 15m30s | 3m06s |
| SWE-bench | $1.92 | $1.44 | 9m27s | 5m50s |
| Terminal-Bench 4.0 | $17.14 | $9.02 | 1h11m | 1h24m |
| Vibe Code Bench v1.1 | $26.88 | $4.83 | 1h16m | 23m43s |
Results available only for Claude Opus 4.8
- Public Benefits Bench v1.1
- Time Horizon Index: KSP
Results available only for Gemini 3.7 Flash
- ProofBench v1.1
- IOI
- CyberBench v1.1
- SRE Bench