Model comparison
Gemini 3.7 Flash vs MiMo V2.6 Flash: Benchmark Comparison
Gemini 3.7 Flash has the higher score on 6 of 18 shared benchmarks; MiMo V2.6 Flash leads on 11.
The largest observed score gap is 31.61 pts on CyberBench v1.1 , where MiMo V2.6 Flash leads.
Reported ±1 standard-error ranges overlap on 12 of 18 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 | MiMo V2.6 Flash | Gap | Reported uncertainty |
|---|---|---|---|---|
| Vals Index | 59.31% ±1.06 | 59.58% ±1.13 | 0.28 pts | Reported ±1 SE ranges overlap |
| Harvey's Legal Agent Benchmark | 8.75% ±2.15 | 11.25% ±2.29 | 2.50 pts | Reported ±1 SE ranges overlap |
| Legal Research Bench | 34.62% ±3.31 | 37.98% ±3.37 | 3.37 pts | Reported ±1 SE ranges overlap |
| EMB | 71.33% ±2.26 | 65.46% ±2.71 | 5.87 pts | Reported ±1 SE ranges do not overlap |
| Finance Agent (v2) | 59.04% ±0.27 | 56.28% ±0.46 | 2.77 pts | Reported ±1 SE ranges do not overlap |
| Tax Agent Bench | 57.66% ±3.31 | 59.90% ±3.13 | 2.24 pts | Reported ±1 SE ranges overlap |
| MedCode | 53.39% ±2.12 | 41.06% ±2.03 | 12.33 pts | Reported ±1 SE ranges do not overlap |
| MedScribe | 83.94% ±2.00 | 85.28% ±1.99 | 1.33 pts | Reported ±1 SE ranges overlap |
| ProofBench v1.1 | 58.00% ±4.96 | 63.00% ±4.85 | 5.00 pts | Reported ±1 SE ranges overlap |
| Terminal-Bench Science | 5.71% ±2.79 | 5.71% ±2.79 | 0.00 pts | Reported ±1 SE ranges overlap |
| SAGE | 49.23% ±3.38 | 43.53% ±3.43 | 5.70 pts | Reported ±1 SE ranges overlap |
| Code Migration | 34.80% ±4.22 | 40.93% ±4.35 | 6.13 pts | Reported ±1 SE ranges overlap |
| IOI | 67.83% ±3.61 | 47.67% ±4.79 | 20.17 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 |
| Terminal-Bench 4.0 | 6.06% ±0.88 | 21.21% ±1.75 | 15.15 pts | Reported ±1 SE ranges do not overlap |
| Vibe Code Bench v1.1 | 70.39% ±4.84 | 78.96% ±4.04 | 8.56 pts | Reported ±1 SE ranges overlap |
| CyberBench v1.1 | 43.75% ±2.21 | 75.36% ±5.42 | 31.61 pts | Reported ±1 SE ranges do not overlap |
| SRE Bench | 4.58% ±1.29 | 4.20% ±1.24 | 0.38 pts | Reported ±1 SE ranges overlap |
Performance by category
| Category | Gemini 3.7 Flash average | MiMo V2.6 Flash average |
|---|---|---|
| Index | 59.31% | 59.58% |
| Legal | 21.68% | 24.62% |
| Finance | 62.68% | 60.55% |
| Healthcare | 68.67% | 63.17% |
| Math | 58.00% | 63.00% |
| Science | 5.71% | 5.71% |
| Education | 49.23% | 43.53% |
| Coding | 35.82% | 37.85% |
| Beta | 24.16% | 39.78% |
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 | MiMo V2.6 Flash cost | Gemini 3.7 Flash latency | MiMo V2.6 Flash latency |
|---|---|---|---|---|
| Vals Index | $4.17 | $0.20 | 43m13s | 45m02s |
| Harvey's Legal Agent Benchmark | $2.56 | $0.09 | 7m32s | 17m12s |
| Legal Research Bench | $0.76 | $0.07 | 4m49s | 15m20s |
| EMB | $5.69 | $0.11 | 10m51s | 25m22s |
| Finance Agent (v2) | $1.48 | $0.07 | 3m01s | 7m02s |
| Tax Agent Bench | $0.48 | $0.05 | 87.24s | 17m19s |
| MedCode | N/A | N/A | 9.16s | 93.37s |
| MedScribe | N/A | N/A | 16.82s | 74.48s |
| ProofBench v1.1 | $0.56 | $0.12 | 7m04s | 1h03m |
| Terminal-Bench Science | $13.22 | $0.31 | 1h08m | 3h37m |
| SAGE | N/A | N/A | 9.79s | 2m16s |
| Code Migration | $21.46 | $0.49 | 2h39m | 3h12m |
| IOI | $3.40 | $0.22 | 21m09s | 1h49m |
| ProgramBench | $7.13 | $0.47 | 21m30s | 3h22m |
| Terminal-Bench 4.0 | $14.07 | $0.22 | 1h59m | 2h21m |
| Vibe Code Bench v1.1 | $4.83 | $0.56 | 23m43s | 50m13s |
| CyberBench v1.1 | $2.15 | $0.05 | 7m34s | 28m22s |
| SRE Bench | $21.33 | $0.40 | 1h50m | 3h22m |
Results available only for Gemini 3.7 Flash
- Vals RSI Index
- LegalBench
- MortgageTax
- TaxEval v2
- GPQA Diamond
- MMLU Pro
- MMMU Pro
- LiveCodeBench
- SkillsBench
- SWE-bench
Results available only for MiMo V2.6 Flash
- BioMysteryBench
- MysteryMechanism
- Public Benefits Bench v1.1