Gemini 3.6 Flash vs GPT 5.5: Benchmark Comparison
Gemini 3.6 Flash has the higher score on 8 of 20 shared benchmarks; GPT 5.5 leads on 12.
The largest observed score gap is 15.38 pts on Legal Research Bench , where GPT 5.5 leads.
Reported ±1 standard-error ranges overlap on 12 of 20 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 3.6 Flash | GPT 5.5 | Gap | Reported uncertainty |
|---|---|---|---|---|
| Harvey's Legal Agent Benchmark | 3.33% ±1.17 | 3.75% ±1.17 | 0.42 pts | Reported ±1 SE ranges overlap |
| Legal Research Bench | 25.00% ±3.01 | 40.38% ±3.41 | 15.38 pts | Reported ±1 SE ranges do not overlap |
| LegalBench | 86.70% ±0.41 | 86.52% ±0.41 | 0.19 pts | Reported ±1 SE ranges overlap |
| EMB | 65.41% ±2.65 | 64.54% ±2.87 | 0.87 pts | Reported ±1 SE ranges overlap |
| Finance Agent (v2) | 56.30% ±0.18 | 51.76% ±0.55 | 4.54 pts | Reported ±1 SE ranges do not overlap |
| MortgageTax | 67.61% ±0.92 | 68.76% ±0.91 | 1.15 pts | Reported ±1 SE ranges overlap |
| Tax Agent Bench | 49.90% ±3.30 | 60.46% ±3.22 | 10.56 pts | Reported ±1 SE ranges do not overlap |
| TaxEval v2 | 74.86% ±0.85 | 74.98% ±0.85 | 0.12 pts | Reported ±1 SE ranges overlap |
| MedCode | 53.15% ±2.16 | 49.10% ±2.19 | 4.05 pts | Reported ±1 SE ranges overlap |
| MedScribe | 79.66% ±1.86 | 86.87% ±1.93 | 7.21 pts | Reported ±1 SE ranges do not overlap |
| GPQA Diamond | 93.43% ±1.33 | 93.18% ±1.29 | 0.25 pts | Reported ±1 SE ranges overlap |
| MMLU Pro | 89.28% ±0.30 | 88.14% ±0.32 | 1.13 pts | Reported ±1 SE ranges do not overlap |
| MMMU Pro | 88.38% ±0.77 | 88.27% ±0.77 | 0.12 pts | Reported ±1 SE ranges overlap |
| SAGE | 48.69% ±3.38 | 51.53% ±3.95 | 2.85 pts | Reported ±1 SE ranges overlap |
| Code Migration | 30.93% ±4.07 | 45.16% ±4.16 | 14.23 pts | Reported ±1 SE ranges do not overlap |
| LiveCodeBench | 88.08% ±0.94 | 85.30% ±1.02 | 2.78 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 |
| SWE-bench | 79.60% ±1.80 | 82.60% ±1.70 | 3.00 pts | Reported ±1 SE ranges overlap |
| Vibe Code Bench v1.1 | 64.00% ±4.29 | 69.85% ±4.54 | 5.84 pts | Reported ±1 SE ranges overlap |
| Public Benefits Bench v1.1 | 56.56% ±1.29 | 60.89% ±1.27 | 4.33 pts | Reported ±1 SE ranges do not overlap |
Performance by category
| Category | Gemini 3.6 Flash average | GPT 5.5 average |
|---|---|---|
| Legal | 38.35% | 43.55% |
| Finance | 62.81% | 64.10% |
| Healthcare | 66.41% | 67.98% |
| Academic | 90.36% | 89.86% |
| Education | 48.69% | 51.53% |
| Coding | 52.52% | 56.68% |
| Social Mobility | 56.56% | 60.89% |
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 3.6 Flash cost | GPT 5.5 cost | Gemini 3.6 Flash latency | GPT 5.5 latency |
|---|---|---|---|---|
| Harvey's Legal Agent Benchmark | $1.90 | $4.60 | 16m24s | 12m14s |
| Legal Research Bench | $0.47 | $7.40 | 2m48s | 37m45s |
| LegalBench | N/A | N/A | 2.96s | 18.14s |
| EMB | $3.89 | $3.27 | 12m06s | 15m11s |
| Finance Agent (v2) | $1.40 | $4.15 | 3m45s | 11m02s |
| MortgageTax | N/A | N/A | 9.24s | 28.24s |
| Tax Agent Bench | $0.36 | $4.07 | 105.32s | 15m52s |
| TaxEval v2 | N/A | N/A | 12.34s | 95.83s |
| MedCode | N/A | N/A | 16.68s | 2m40s |
| MedScribe | N/A | N/A | 33.55s | 2m13s |
| GPQA Diamond | N/A | N/A | 18.06s | 112.33s |
| MMLU Pro | N/A | N/A | 7.30s | 42.11s |
| MMMU Pro | N/A | N/A | 11.31s | 54.15s |
| SAGE | N/A | N/A | 21.96s | 76.05s |
| Code Migration | $8.93 | $6.44 | 1h30m | 31m35s |
| LiveCodeBench | N/A | N/A | 31.88s | 2m47s |
| ProgramBench | $5.92 | $6.95 | 47m33s | 22m03s |
| SWE-bench | $1.19 | $1.36 | 13m08s | 7m06s |
| Vibe Code Bench v1.1 | $3.04 | $16.66 | 25m23s | 31m52s |
| Public Benefits Bench v1.1 | $0.39 | $3.97 | 2m02s | 39m37s |
Results available only for Gemini 3.6 Flash
- BioMysteryBench
- Terminal-Bench Science
- IOI
- CyberBench v1.1
Results available only for GPT 5.5
- Vals RSI Index
- SkillsBench
- SRE Bench
- Time Horizon Index: KSP