Gemini 3.7 Flash vs Kimi K3: Benchmark Comparison
Gemini 3.7 Flash has the higher score on 9 of 20 shared benchmarks; Kimi K3 leads on 11.
The largest observed score gap is 31.49 pts on CyberBench v1.1 , where Kimi K3 leads.
Reported ±1 standard-error ranges overlap on 6 of 19 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 | Kimi K3 | Gap | Reported uncertainty |
|---|---|---|---|---|
| Vals RSI Index | 18.27% | 20.88% | 2.61 pts | Uncertainty comparison unavailable |
| Legal Research Bench | 34.62% ±3.31 | 46.15% ±3.46 | 11.54 pts | Reported ±1 SE ranges do not overlap |
| LegalBench | 87.26% ±0.42 | 86.21% ±0.41 | 1.04 pts | Reported ±1 SE ranges do not overlap |
| MortgageTax | 66.65% ±0.92 | 66.34% ±0.91 | 0.32 pts | Reported ±1 SE ranges overlap |
| Tax Agent Bench | 22.80% ±3.03 | 33.37% ±3.39 | 10.57 pts | Reported ±1 SE ranges do not overlap |
| TaxEval v2 | 74.73% ±0.85 | 75.72% ±0.84 | 0.98 pts | Reported ±1 SE ranges overlap |
| MedCode | 53.39% ±2.12 | 49.36% ±2.20 | 4.03 pts | Reported ±1 SE ranges overlap |
| MedScribe | 83.94% ±2.00 | 88.05% ±1.98 | 4.11 pts | Reported ±1 SE ranges do not overlap |
| ProofBench v1.1 | 58.00% ±4.96 | 87.00% ±3.38 | 29.00 pts | Reported ±1 SE ranges do not overlap |
| Terminal-Bench-Science | 7.14% ±3.10 | 1.43% ±1.43 | 5.71 pts | Reported ±1 SE ranges do not overlap |
| GPQA Diamond | 93.94% ±1.49 | 92.93% ±1.31 | 1.01 pts | Reported ±1 SE ranges overlap |
| MMLU Pro | 90.12% ±0.30 | 87.97% ±0.32 | 2.15 pts | Reported ±1 SE ranges do not overlap |
| MMMU Pro | 88.96% ±0.75 | 88.15% ±0.78 | 0.81 pts | Reported ±1 SE ranges overlap |
| IOI | 67.83% ±3.61 | 48.94% ±9.82 | 18.89 pts | Reported ±1 SE ranges do not overlap |
| LiveCodeBench | 88.65% ±0.92 | 87.19% ±0.97 | 1.47 pts | Reported ±1 SE ranges overlap |
| ProgramBench | 0.00% ±0.00 | 2.00% ±0.99 | 2.00 pts | Reported ±1 SE ranges do not overlap |
| SWE-bench | 80.80% ±1.76 | 93.40% ±1.11 | 12.60 pts | Reported ±1 SE ranges do not overlap |
| Terminal-Bench 4.0 | 12.12% ±3.03 | 17.17% ±0.51 | 5.05 pts | Reported ±1 SE ranges do not overlap |
| Vibe Code Bench v1.1 | 70.39% ±4.84 | 84.97% ±2.75 | 14.57 pts | Reported ±1 SE ranges do not overlap |
| CyberBench v1.1 | 43.75% ±2.21 | 75.24% ±5.56 | 31.49 pts | Reported ±1 SE ranges do not overlap |
Performance by category
| Category | Gemini 3.7 Flash average | Kimi K3 average |
|---|---|---|
| Legal | 60.94% | 66.18% |
| Finance | 54.73% | 58.47% |
| Healthcare | 68.67% | 68.70% |
| Math | 58.00% | 87.00% |
| Science | 7.14% | 1.43% |
| Academic | 91.01% | 89.68% |
| Coding | 53.30% | 55.61% |
| Cyber | 43.75% | 75.24% |
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 | Kimi K3 cost | Gemini 3.7 Flash latency | Kimi K3 latency |
|---|---|---|---|---|
| Vals RSI Index | $180.16 | $136.82 | 90h00m | 90h00m |
| Legal Research Bench | $0.76 | $3.47 | 4m49s | 12m40s |
| LegalBench | N/A | N/A | 2.23s | 7.11s |
| MortgageTax | N/A | N/A | 4.11s | 78.94s |
| Tax Agent Bench | $0.48 | $2.79 | 87.24s | 40m15s |
| TaxEval v2 | N/A | N/A | 7.83s | 2m25s |
| MedCode | N/A | N/A | 9.16s | 36.99s |
| MedScribe | N/A | N/A | 16.82s | 38.24s |
| ProofBench v1.1 | $0.56 | $1.65 | 7m04s | 34m31s |
| Terminal-Bench-Science | $7.19 | $21.30 | 1h20m | 4h49m |
| GPQA Diamond | N/A | N/A | 8.77s | 2m02s |
| MMLU Pro | N/A | N/A | 4.16s | 38.72s |
| MMMU Pro | N/A | N/A | 6.28s | 106.52s |
| IOI | $3.40 | $14.17 | 21m09s | 4h18m |
| LiveCodeBench | N/A | N/A | 14.53s | 3m20s |
| ProgramBench | $7.13 | $70.48 | 21m30s | 5h42m |
| SWE-bench | $1.44 | $0.76 | 5m50s | 10m19s |
| Terminal-Bench 4.0 | $8.71 | $12.48 | 1h24m | 3h09m |
| Vibe Code Bench v1.1 | $4.83 | $10.01 | 23m43s | 16m39s |
| CyberBench v1.1 | $2.15 | $2.13 | 7m34s | 30m44s |
Results available only for Gemini 3.7 Flash
- SkillsBench
- SRE Bench
Results available only for Kimi K3
- Vals Index
- Vals Finance Index
- Vals Legal Index
- Vals Inference Provider Index
- HLAB-Vals-1.1
- Immigration
- EMB v1.1
- Finance Agent v2.1
- BioMysteryBench
- SAGE
- Code Migration v1.1
- SWE-bench Multimodal
- Vibe Code Bench 1-100
- CyberOffense
- Access to Justice Benchmark
- Public Benefits Bench v1.1
- Time Horizon Index: KSP