Model comparison
Gemini 3.7 Flash vs GPT 5.5: Benchmark Comparison
Gemini 3.7 Flash has the higher score on 14 of 23 shared benchmarks; GPT 5.5 leads on 9.
The largest observed score gap is 10.36 pts on Code Migration , where GPT 5.5 leads.
Reported ±1 standard-error ranges overlap on 15 of 22 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 | Gemini 3.7 Flash | GPT 5.5 | Gap | Reported uncertainty |
|---|---|---|---|---|
| Vals Index | 59.31% ±1.06 | 57.41% ±1.15 | 1.90 pts | Reported ±1 SE ranges overlap |
| Vals RSI Index | 18.27% | 16.48% | 1.79 pts | Uncertainty comparison unavailable |
| Harvey's Legal Agent Benchmark | 8.75% ±2.15 | 3.75% ±1.17 | 5.00 pts | Reported ±1 SE ranges do not overlap |
| Legal Research Bench | 34.62% ±3.31 | 40.38% ±3.41 | 5.77 pts | Reported ±1 SE ranges overlap |
| LegalBench | 87.26% ±0.42 | 86.52% ±0.41 | 0.74 pts | Reported ±1 SE ranges overlap |
| EMB | 71.33% ±2.26 | 64.54% ±2.87 | 6.78 pts | Reported ±1 SE ranges do not overlap |
| Finance Agent (v2) | 59.04% ±0.27 | 51.76% ±0.55 | 7.28 pts | Reported ±1 SE ranges do not overlap |
| MortgageTax | 66.65% ±0.92 | 68.76% ±0.91 | 2.11 pts | Reported ±1 SE ranges do not overlap |
| Tax Agent Bench | 57.66% ±3.31 | 60.46% ±3.22 | 2.79 pts | Reported ±1 SE ranges overlap |
| TaxEval v2 | 74.73% ±0.85 | 74.98% ±0.85 | 0.25 pts | Reported ±1 SE ranges overlap |
| MedCode | 53.39% ±2.12 | 49.10% ±2.19 | 4.29 pts | Reported ±1 SE ranges overlap |
| MedScribe | 83.94% ±2.00 | 86.87% ±1.93 | 2.93 pts | Reported ±1 SE ranges overlap |
| GPQA Diamond | 93.94% ±1.49 | 93.18% ±1.29 | 0.76 pts | Reported ±1 SE ranges overlap |
| MMLU Pro | 90.12% ±0.30 | 88.14% ±0.32 | 1.98 pts | Reported ±1 SE ranges do not overlap |
| MMMU Pro | 88.96% ±0.75 | 88.27% ±0.77 | 0.69 pts | Reported ±1 SE ranges overlap |
| SAGE | 49.23% ±3.38 | 51.53% ±3.95 | 2.30 pts | Reported ±1 SE ranges overlap |
| Code Migration | 34.80% ±4.22 | 45.16% ±4.16 | 10.36 pts | Reported ±1 SE ranges do not overlap |
| LiveCodeBench | 88.65% ±0.92 | 85.30% ±1.02 | 3.36 pts | Reported ±1 SE ranges do not overlap |
| ProgramBench | 0.00% ±0.00 | 0.50% ±0.50 | 0.50 pts | Reported ±1 SE ranges overlap |
| SkillsBench | 65.89% ±4.55 | 62.21% ±4.52 | 3.68 pts | Reported ±1 SE ranges overlap |
| SWE-bench | 80.80% ±1.76 | 82.60% ±1.70 | 1.80 pts | Reported ±1 SE ranges overlap |
| Vibe Code Bench v1.1 | 70.39% ±4.84 | 69.85% ±4.54 | 0.55 pts | Reported ±1 SE ranges overlap |
| SRE Bench | 4.58% ±1.29 | 3.82% ±1.19 | 0.76 pts | Reported ±1 SE ranges overlap |
Performance by category
| Category | Gemini 3.7 Flash average | GPT 5.5 average |
|---|---|---|
| Index | 38.79% | 36.95% |
| Legal | 43.54% | 43.55% |
| Finance | 65.88% | 64.10% |
| Healthcare | 68.67% | 67.98% |
| Academic | 91.01% | 89.86% |
| Education | 49.23% | 51.53% |
| Coding | 56.76% | 57.60% |
| Beta | 4.58% | 3.82% |
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 | Gemini 3.7 Flash cost | GPT 5.5 cost | Gemini 3.7 Flash latency | GPT 5.5 latency |
|---|---|---|---|---|
| Vals Index | $4.17 | $6.14 | 43m13s | 20m51s |
| Vals RSI Index | $180.16 | $857.16 | 90h00m | 90h00m |
| Harvey's Legal Agent Benchmark | $2.56 | $4.60 | 7m32s | 12m14s |
| Legal Research Bench | $0.76 | $7.40 | 4m49s | 37m45s |
| LegalBench | N/A | N/A | 2.23s | 18.14s |
| EMB | $5.69 | $3.27 | 10m51s | 15m11s |
| Finance Agent (v2) | $1.48 | $4.15 | 3m01s | 11m02s |
| MortgageTax | N/A | N/A | 4.11s | 28.24s |
| Tax Agent Bench | $0.48 | $4.07 | 87.24s | 15m52s |
| TaxEval v2 | N/A | N/A | 7.83s | 95.83s |
| MedCode | N/A | N/A | 9.16s | 2m40s |
| MedScribe | N/A | N/A | 16.82s | 2m13s |
| GPQA Diamond | N/A | N/A | 8.77s | 112.33s |
| MMLU Pro | N/A | N/A | 4.16s | 42.11s |
| MMMU Pro | N/A | N/A | 6.28s | 54.15s |
| SAGE | N/A | N/A | 9.79s | 76.05s |
| Code Migration | $21.46 | $6.44 | 2h39m | 31m35s |
| LiveCodeBench | N/A | N/A | 14.53s | 2m47s |
| ProgramBench | $7.13 | $6.95 | 21m30s | 22m03s |
| SkillsBench | $1.80 | $2.54 | 3m06s | 6m55s |
| SWE-bench | $1.44 | $1.36 | 5m50s | 7m06s |
| Vibe Code Bench v1.1 | $4.83 | $16.66 | 23m43s | 31m52s |
| SRE Bench | $21.33 | $17.77 | 1h50m | 46m28s |
Results available only for Gemini 3.7 Flash
- ProofBench v1.1
- Terminal-Bench Science
- IOI
- Terminal-Bench 4.0
- CyberBench v1.1
Results available only for GPT 5.5
- Time Horizon Index: KSP
- Public Benefits Bench v1.1