Gemini 3.5 Flash vs Muse Spark 1.2: Benchmark Comparison
Gemini 3.5 Flash has the higher score on 5 of 17 shared benchmarks; Muse Spark 1.2 leads on 11.
The largest observed score gap is 30.42 pts on Vibe Code Bench v1.1 , where Muse Spark 1.2 leads.
Reported ±1 standard-error ranges overlap on 4 of 17 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 3.5 Flash | Muse Spark 1.2 | Gap | Reported uncertainty |
|---|---|---|---|---|
| Legal Research Bench | 30.77% ±3.21 | 43.75% ±3.45 | 12.98 pts | Reported ±1 SE ranges do not overlap |
| LegalBench | 83.60% ±0.85 | 85.26% ±0.45 | 1.66 pts | Reported ±1 SE ranges do not overlap |
| MortgageTax | 68.12% ±0.91 | 65.42% ±0.93 | 2.70 pts | Reported ±1 SE ranges do not overlap |
| Tax Agent Bench | 21.65% ±2.96 | 32.80% ±2.75 | 11.15 pts | Reported ±1 SE ranges do not overlap |
| TaxEval v2 | 74.37% ±0.85 | 80.38% ±0.76 | 6.01 pts | Reported ±1 SE ranges do not overlap |
| MedCode | 55.83% ±2.11 | 49.35% ±2.19 | 6.48 pts | Reported ±1 SE ranges do not overlap |
| MedScribe | 76.57% ±1.92 | 90.06% ±1.96 | 13.49 pts | Reported ±1 SE ranges do not overlap |
| ProofBench v1.1 | 31.00% ±4.65 | 43.00% ±4.98 | 12.00 pts | Reported ±1 SE ranges do not overlap |
| MMLU Pro | 89.52% ±0.31 | 88.28% ±0.32 | 1.24 pts | Reported ±1 SE ranges do not overlap |
| MMMU Pro | 88.27% ±0.77 | 86.13% ±0.83 | 2.14 pts | Reported ±1 SE ranges do not overlap |
| SAGE | 49.88% ±3.44 | 47.66% ±3.44 | 2.23 pts | Reported ±1 SE ranges overlap |
| ProgramBench | 0.00% ±0.00 | 0.50% ±0.50 | 0.50 pts | Reported ±1 SE ranges overlap |
| SkillsBench | 52.74% ±4.47 | 53.04% ±4.42 | 0.30 pts | Reported ±1 SE ranges overlap |
| SWE-bench | 78.80% ±1.83 | 86.60% ±1.52 | 7.80 pts | Reported ±1 SE ranges do not overlap |
| Terminal-Bench 4.0 | 6.06% ±1.51 | 6.06% ±1.51 | 0.00 pts | Reported ±1 SE ranges overlap |
| Vibe Code Bench v1.1 | 48.68% ±4.73 | 79.10% ±3.31 | 30.42 pts | Reported ±1 SE ranges do not overlap |
| Public Benefits Bench v1.1 | 59.47% ±1.28 | 68.47% ±1.21 | 9.00 pts | Reported ±1 SE ranges do not overlap |
Performance by category
| Category | Gemini 3.5 Flash average | Muse Spark 1.2 average |
|---|---|---|
| Legal | 57.19% | 64.51% |
| Finance | 54.71% | 59.53% |
| Healthcare | 66.20% | 69.70% |
| Math | 31.00% | 43.00% |
| Academic | 88.89% | 87.20% |
| Education | 49.88% | 47.66% |
| Coding | 37.26% | 45.06% |
| Social Mobility | 59.47% | 68.47% |
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 3.5 Flash cost | Muse Spark 1.2 cost | Gemini 3.5 Flash latency | Muse Spark 1.2 latency |
|---|---|---|---|---|
| Legal Research Bench | $1.31 | $0.51 | 4m38s | 5m07s |
| LegalBench | N/A | N/A | 34.58s | 30.10s |
| MortgageTax | N/A | N/A | 11.43s | 44.62s |
| Tax Agent Bench | $0.86 | $0.27 | 3m53s | 119.91s |
| TaxEval v2 | N/A | N/A | 19.69s | 48.31s |
| MedCode | N/A | N/A | 25.29s | 60.17s |
| MedScribe | N/A | N/A | 57.08s | 61.46s |
| ProofBench v1.1 | $0.57 | $0.44 | 6m34s | 7m49s |
| MMLU Pro | N/A | N/A | 26.26s | 44.23s |
| MMMU Pro | N/A | N/A | 12.23s | 73.84s |
| SAGE | N/A | N/A | 32.59s | 29.07s |
| ProgramBench | N/A | N/A | 18m25s | 26m38s |
| SkillsBench | $1.25 | $0.65 | 7m15s | 13m00s |
| SWE-bench | $0.95 | $0.55 | 4m14s | 8m07s |
| Terminal-Bench 4.0 | $5.98 | $4.14 | 1h11m | 1h18m |
| Vibe Code Bench v1.1 | $2.54 | $1.53 | 14m53s | 19m55s |
| Public Benefits Bench v1.1 | $1.11 | $0.59 | 5m39s | 8m55s |
Results available only for Gemini 3.5 Flash
- Terminal-Bench-Science
- GPQA Diamond
- LiveCodeBench
Results available only for Muse Spark 1.2
- BioMysteryBench
- IOI
- CyberBench v1.1
- SRE Bench