Model comparison
Gemini 3.7 Flash vs Grok 4.6: Benchmark Comparison
Gemini 3.7 Flash has the higher score on 14 of 25 shared benchmarks; Grok 4.6 leads on 11.
The largest observed score gap is 23.93 pts on CyberBench v1.1 , where Grok 4.6 leads.
Reported ±1 standard-error ranges overlap on 7 of 24 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 | Grok 4.6 | Gap | Reported uncertainty |
|---|---|---|---|---|
| Vals Index | 59.31% ±1.06 | 59.17% ±1.21 | 0.14 pts | Reported ±1 SE ranges overlap |
| Vals RSI Index | 18.27% | 25.46% | 7.19 pts | Uncertainty comparison unavailable |
| Harvey's Legal Agent Benchmark | 8.75% ±2.15 | 15.83% ±2.65 | 7.08 pts | Reported ±1 SE ranges do not overlap |
| Legal Research Bench | 34.62% ±3.31 | 48.08% ±3.47 | 13.46 pts | Reported ±1 SE ranges do not overlap |
| LegalBench | 87.26% ±0.42 | 86.31% ±0.42 | 0.95 pts | Reported ±1 SE ranges do not overlap |
| EMB | 71.33% ±2.26 | 62.73% ±3.08 | 8.60 pts | Reported ±1 SE ranges do not overlap |
| Finance Agent (v2) | 59.04% ±0.27 | 53.68% ±0.67 | 5.36 pts | Reported ±1 SE ranges do not overlap |
| MortgageTax | 66.65% ±0.92 | 64.19% ±0.95 | 2.47 pts | Reported ±1 SE ranges do not overlap |
| Tax Agent Bench | 57.66% ±3.31 | 70.79% ±3.00 | 13.12 pts | Reported ±1 SE ranges do not overlap |
| TaxEval v2 | 74.73% ±0.85 | 71.10% ±0.90 | 3.64 pts | Reported ±1 SE ranges do not overlap |
| MedCode | 53.39% ±2.12 | 44.71% ±2.26 | 8.68 pts | Reported ±1 SE ranges do not overlap |
| MedScribe | 83.94% ±2.00 | 86.53% ±1.96 | 2.59 pts | Reported ±1 SE ranges overlap |
| ProofBench v1.1 | 58.00% ±4.96 | 51.00% ±5.02 | 7.00 pts | Reported ±1 SE ranges overlap |
| Terminal-Bench Science | 5.71% ±2.79 | 4.29% ±2.44 | 1.43 pts | Reported ±1 SE ranges overlap |
| GPQA Diamond | 93.94% ±1.49 | 94.70% ±1.13 | 0.76 pts | Reported ±1 SE ranges overlap |
| MMLU Pro | 90.12% ±0.30 | 89.40% ±0.30 | 0.72 pts | Reported ±1 SE ranges do not overlap |
| SAGE | 49.23% ±3.38 | 28.90% ±3.08 | 20.33 pts | Reported ±1 SE ranges do not overlap |
| Code Migration | 34.80% ±4.22 | 44.57% ±4.48 | 9.77 pts | Reported ±1 SE ranges do not overlap |
| IOI | 67.83% ±3.61 | 47.61% ±2.09 | 20.22 pts | Reported ±1 SE ranges do not overlap |
| LiveCodeBench | 88.65% ±0.92 | 88.22% ±0.94 | 0.43 pts | Reported ±1 SE ranges overlap |
| SkillsBench | 65.89% ±4.55 | 55.77% ±4.84 | 10.13 pts | Reported ±1 SE ranges do not overlap |
| SWE-bench | 80.80% ±1.76 | 95.60% ±0.92 | 14.80 pts | Reported ±1 SE ranges do not overlap |
| Terminal-Bench 4.0 | 6.06% ±0.88 | 17.17% ±2.81 | 11.11 pts | Reported ±1 SE ranges do not overlap |
| Vibe Code Bench v1.1 | 70.39% ±4.84 | 76.24% ±3.82 | 5.84 pts | Reported ±1 SE ranges overlap |
| CyberBench v1.1 | 43.75% ±2.21 | 67.68% ±5.87 | 23.93 pts | Reported ±1 SE ranges do not overlap |
Performance by category
| Category | Gemini 3.7 Flash average | Grok 4.6 average |
|---|---|---|
| Index | 38.79% | 42.31% |
| Legal | 43.54% | 50.07% |
| Finance | 65.88% | 64.50% |
| Healthcare | 68.67% | 65.62% |
| Math | 58.00% | 51.00% |
| Science | 5.71% | 4.29% |
| Academic | 92.03% | 92.05% |
| Education | 49.23% | 28.90% |
| Coding | 59.20% | 60.74% |
| Beta | 43.75% | 67.68% |
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 | Grok 4.6 cost | Gemini 3.7 Flash latency | Grok 4.6 latency |
|---|---|---|---|---|
| Vals Index | $4.17 | $4.34 | 43m13s | 38m04s |
| Vals RSI Index | $180.16 | $88.12 | 90h00m | 90h00m |
| Harvey's Legal Agent Benchmark | $2.56 | $4.01 | 7m32s | 45m05s |
| Legal Research Bench | $0.76 | $1.53 | 4m49s | 25m04s |
| LegalBench | N/A | N/A | 2.23s | 27.42s |
| EMB | $5.69 | $3.06 | 10m51s | 52m59s |
| Finance Agent (v2) | $1.48 | $1.66 | 3m01s | 16m41s |
| MortgageTax | N/A | N/A | 4.11s | 8.55s |
| Tax Agent Bench | $0.48 | $0.98 | 87.24s | 10m58s |
| TaxEval v2 | N/A | N/A | 7.83s | 47.57s |
| MedCode | N/A | N/A | 9.16s | 2m07s |
| MedScribe | N/A | N/A | 16.82s | 70.94s |
| ProofBench v1.1 | $0.56 | $0.76 | 7m04s | 14m22s |
| Terminal-Bench Science | $13.22 | $10.70 | 1h08m | 2h32m |
| GPQA Diamond | N/A | N/A | 8.77s | 3m29s |
| MMLU Pro | N/A | N/A | 4.16s | 54.98s |
| SAGE | N/A | N/A | 9.79s | 3m22s |
| Code Migration | $21.46 | $14.58 | 2h39m | 1h30m |
| IOI | $3.40 | $8.18 | 21m09s | 3h57m |
| LiveCodeBench | N/A | N/A | 14.53s | 118.40s |
| SkillsBench | $1.80 | $0.92 | 3m06s | 13m28s |
| SWE-bench | $1.44 | $0.78 | 5m50s | 10m04s |
| Terminal-Bench 4.0 | $14.07 | $6.47 | 1h59m | 3h15m |
| Vibe Code Bench v1.1 | $4.83 | $4.88 | 23m43s | 25m30s |
| CyberBench v1.1 | $2.15 | $4.07 | 7m34s | 36m52s |
Results available only for Gemini 3.7 Flash
- MMMU Pro
- ProgramBench
- SRE Bench
Results available only for Grok 4.6
- BioMysteryBench
- MysteryMechanism
- Vibe Code Bench 1-100
- Time Horizon Index: KSP
- Public Benefits Bench v1.1