Gemini 3.8 Flash vs Muse Spark 1.3 Max: Benchmark Comparison
Gemini 3.8 Flash has the higher score on 5 of 18 shared benchmarks; Muse Spark 1.3 Max leads on 13.
The largest observed score gap is 28.99 pts on CyberBench v1.1 , where Muse Spark 1.3 Max leads.
Reported ±1 standard-error ranges overlap on 8 of 16 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 | Muse Spark 1.3 Max | Gap | Reported uncertainty |
|---|---|---|---|---|
| Vals Index | 54.83% ±1.03 | 58.16% ±1.19 | 3.34 pts | Reported ±1 SE ranges do not overlap |
| Vals RSI Index | 19.93% | 19.64% | 0.29 pts | Uncertainty comparison unavailable |
| Harvey's Legal Agent Benchmark | 10.00% ±2.42 | 23.75% ±3.55 | 13.75 pts | Reported ±1 SE ranges do not overlap |
| Legal Research Bench | 38.94% ±3.39 | 55.29% ±3.46 | 16.35 pts | Reported ±1 SE ranges do not overlap |
| EMB | 72.20% ±2.42 | 67.43% ±3.06 | 4.77 pts | Reported ±1 SE ranges overlap |
| Finance Agent (v2) | 61.44% ±0.13 | 59.96% ±2.06 | 1.48 pts | Reported ±1 SE ranges overlap |
| Tax Agent Bench | 66.77% ±3.15 | 72.44% ±2.88 | 5.67 pts | Reported ±1 SE ranges overlap |
| ProofBench v1.1 | 48.00% ±5.02 | 58.00% ±4.96 | 10.00 pts | Reported ±1 SE ranges do not overlap |
| MysteryMechanism | 36.49% ±3.24 | 36.04% ±3.23 | 0.45 pts | Reported ±1 SE ranges overlap |
| Terminal-Bench Science | 8.57% ±3.37 | 10.00% ±3.61 | 1.43 pts | Reported ±1 SE ranges overlap |
| Code Migration | 36.55% ±4.18 | 47.41% ±4.26 | 10.87 pts | Reported ±1 SE ranges do not overlap |
| IOI | 56.94% ±2.26 | 56.56% ±2.52 | 0.39 pts | Reported ±1 SE ranges overlap |
| ProgramBench | 1.00% ±0.70 | 2.50% ±1.11 | 1.50 pts | Reported ±1 SE ranges overlap |
| Terminal-Bench 4.0 | 19.19% ±2.52 | 24.75% ±0.51 | 5.55 pts | Reported ±1 SE ranges do not overlap |
| Vibe Code Bench 1-100 | 18.77% ±3.31 | 20.46% ±3.85 | 1.69 pts | Reported ±1 SE ranges overlap |
| Vibe Code Bench v1.1 | 78.65% ±3.88 | 85.86% ±2.51 | 7.20 pts | Reported ±1 SE ranges do not overlap |
| CyberBench v1.1 | 43.75% ±2.21 | 72.74% ±5.67 | 28.99 pts | Reported ±1 SE ranges do not overlap |
| CUA-bench | 4.17% | 5.83% | 1.67 pts | Uncertainty comparison unavailable |
Performance by category
| Category | Gemini 3.8 Flash average | Muse Spark 1.3 Max average |
|---|---|---|
| Legal | 24.47% | 39.52% |
| Finance | 66.80% | 66.61% |
| Math | 48.00% | 58.00% |
| Science | 22.53% | 23.02% |
| Coding | 35.18% | 39.59% |
| Cyber | 43.75% | 72.74% |
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 | Muse Spark 1.3 Max cost | Gemini 3.8 Flash latency | Muse Spark 1.3 Max latency |
|---|---|---|---|---|
| Vals Index | $5.73 | $3.79 | 51m57s | 23m33s |
| Vals RSI Index | $332.04 | $48.82 | 90h00m | 90h00m |
| Harvey's Legal Agent Benchmark | $3.66 | $2.26 | 29m21s | 14m35s |
| Legal Research Bench | $1.63 | $0.59 | 5m24s | 5m17s |
| EMB | $8.24 | $2.55 | 12m49s | 13m50s |
| Finance Agent (v2) | $2.00 | $0.76 | 3m21s | 3m19s |
| Tax Agent Bench | $0.86 | $0.39 | 2m58s | 4m09s |
| ProofBench v1.1 | $0.60 | $0.48 | 6m09s | 9m29s |
| MysteryMechanism | $1.73 | $0.85 | 5m22s | 7m32s |
| Terminal-Bench Science | $5.64 | $6.09 | 55m11s | 1h24m |
| Code Migration | $18.49 | $14.21 | 2h37m | 1h24m |
| IOI | $3.98 | $1.73 | 14m20s | 28m58s |
| ProgramBench | $10.93 | $54.29 | 46m39s | 5h28m |
| Terminal-Bench 4.0 | $8.77 | $6.65 | 1h48m | 52m56s |
| Vibe Code Bench 1-100 | $18.91 | $5.60 | 1h06m | 20m02s |
| Vibe Code Bench v1.1 | $6.87 | $2.54 | 8m39s | 16m18s |
| CyberBench v1.1 | $1.31 | $3.34 | 9m49s | 16m18s |
| CUA-bench | $280.03 | $143.48 | 36.55s | 36.39s |
Results available only for Gemini 3.8 Flash
- LegalBench
- MortgageTax
- TaxEval v2
- MedCode
- MedScribe
- BioMysteryBench
- GPQA Diamond
- MMLU Pro
- MMMU Pro
- SAGE
- LiveCodeBench
- SkillsBench
- SWE-bench
- Public Benefits Bench v1.1
- Time Horizon Index: KSP
Results available only for Muse Spark 1.3 Max
None.