Gemini 3.8 Flash vs Grok 4.7: Benchmark Comparison
Gemini 3.8 Flash has the higher score on 7 of 22 shared benchmarks; Grok 4.7 leads on 15.
The largest observed score gap is 25.71 pts on CyberBench v1.1 , where Grok 4.7 leads.
Reported ±1 standard-error ranges overlap on 11 of 21 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 3.8 Flash | Grok 4.7 | Gap | Reported uncertainty |
|---|---|---|---|---|
| Vals Index | 54.83% ±1.03 | 54.95% ±1.07 | 0.12 pts | Reported ±1 SE ranges overlap |
| Vals RSI Index | 19.93% | 20.20% | 0.27 pts | Uncertainty comparison unavailable |
| Harvey's Legal Agent Benchmark | 10.00% ±2.42 | 12.50% ±2.66 | 2.50 pts | Reported ±1 SE ranges overlap |
| Legal Research Bench | 38.94% ±3.39 | 47.12% ±3.47 | 8.17 pts | Reported ±1 SE ranges do not overlap |
| LegalBench | 86.99% ±0.43 | 84.39% ±0.46 | 2.60 pts | Reported ±1 SE ranges do not overlap |
| EMB | 72.20% ±2.42 | 66.99% ±3.04 | 5.20 pts | Reported ±1 SE ranges overlap |
| Finance Agent (v2) | 61.44% ±0.13 | 52.25% ±0.35 | 9.18 pts | Reported ±1 SE ranges do not overlap |
| Tax Agent Bench | 66.77% ±3.15 | 65.60% ±3.24 | 1.17 pts | Reported ±1 SE ranges overlap |
| MedCode | 48.13% ±2.18 | 49.55% ±2.17 | 1.42 pts | Reported ±1 SE ranges overlap |
| MedScribe | 84.50% ±1.94 | 89.38% ±1.89 | 4.88 pts | Reported ±1 SE ranges do not overlap |
| ProofBench v1.1 | 48.00% ±5.02 | 26.00% ±4.41 | 22.00 pts | Reported ±1 SE ranges do not overlap |
| BioMysteryBench | 62.22% ±1.70 | 69.26% ±1.33 | 7.04 pts | Reported ±1 SE ranges do not overlap |
| MysteryMechanism | 36.49% ±3.24 | 25.23% ±2.92 | 11.26 pts | Reported ±1 SE ranges do not overlap |
| Terminal-Bench Science | 8.57% ±3.37 | 10.00% ±3.61 | 1.43 pts | Reported ±1 SE ranges overlap |
| SAGE | 35.06% ±3.36 | 40.79% ±3.35 | 5.73 pts | Reported ±1 SE ranges overlap |
| Code Migration | 36.55% ±4.18 | 44.82% ±4.21 | 8.27 pts | Reported ±1 SE ranges overlap |
| IOI | 56.94% ±2.26 | 57.72% ±1.99 | 0.78 pts | Reported ±1 SE ranges overlap |
| ProgramBench | 1.00% ±0.70 | 0.50% ±0.50 | 0.50 pts | Reported ±1 SE ranges overlap |
| Terminal-Bench 4.0 | 19.19% ±2.52 | 28.79% ±2.31 | 9.60 pts | Reported ±1 SE ranges do not overlap |
| Vibe Code Bench v1.1 | 78.65% ±3.88 | 86.17% ±2.18 | 7.52 pts | Reported ±1 SE ranges do not overlap |
| CyberBench v1.1 | 43.75% ±2.21 | 69.46% ±5.67 | 25.71 pts | Reported ±1 SE ranges do not overlap |
| Public Benefits Bench v1.1 | 65.29% ±1.24 | 65.63% ±1.24 | 0.34 pts | Reported ±1 SE ranges overlap |
Performance by category
| Category | Gemini 3.8 Flash average | Grok 4.7 average |
|---|---|---|
| Legal | 45.31% | 48.00% |
| Finance | 66.80% | 61.61% |
| Healthcare | 66.32% | 69.47% |
| Math | 48.00% | 26.00% |
| Science | 35.76% | 34.83% |
| Education | 35.06% | 40.79% |
| Coding | 38.47% | 43.60% |
| Cyber | 43.75% | 69.46% |
| Social Mobility | 65.29% | 65.63% |
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 3.8 Flash cost | Grok 4.7 cost | Gemini 3.8 Flash latency | Grok 4.7 latency |
|---|---|---|---|---|
| Vals Index | $5.73 | $12.12 | 51m57s | 35m31s |
| Vals RSI Index | $332.04 | $237.12 | 90h00m | 90h00m |
| Harvey's Legal Agent Benchmark | $3.66 | $11.13 | 29m21s | 42m05s |
| Legal Research Bench | $1.63 | $4.92 | 5m24s | 21m14s |
| LegalBench | N/A | N/A | 3.32s | 26.02s |
| EMB | $8.24 | $6.48 | 12m49s | 34m57s |
| Finance Agent (v2) | $2.00 | $2.72 | 3m21s | 15m33s |
| Tax Agent Bench | $0.86 | $1.80 | 2m58s | 16m36s |
| MedCode | N/A | N/A | 42.89s | 3m33s |
| MedScribe | N/A | N/A | 21.14s | 2m33s |
| ProofBench v1.1 | $0.60 | $0.79 | 6m09s | 12m47s |
| BioMysteryBench | $1.55 | $2.49 | 5m27s | 11m37s |
| MysteryMechanism | $1.73 | $1.84 | 5m22s | 17m33s |
| Terminal-Bench Science | $5.64 | $14.45 | 55m11s | 1h29m |
| SAGE | N/A | N/A | 28.64s | 4m51s |
| Code Migration | $18.49 | $36.55 | 2h37m | 1h09m |
| IOI | $3.98 | $12.71 | 14m20s | 46m48s |
| ProgramBench | $10.93 | $46.49 | 46m39s | 3h38m |
| Terminal-Bench 4.0 | $8.77 | $18.09 | 1h48m | 50m21s |
| Vibe Code Bench v1.1 | $6.87 | $15.83 | 8m39s | 35m47s |
| CyberBench v1.1 | $1.31 | $8.54 | 9m49s | 15m30s |
| Public Benefits Bench v1.1 | $1.01 | $2.25 | 3m21s | 16m57s |
Results available only for Gemini 3.8 Flash
- MortgageTax
- TaxEval v2
- GPQA Diamond
- MMLU Pro
- MMMU Pro
- LiveCodeBench
- SkillsBench
- SWE-bench
- Vibe Code Bench 1-100
- CUA-bench
- Time Horizon Index: KSP
Results available only for Grok 4.7
None.