DeepSeek V4 Flash 0731 vs Muse Spark 1.2: Benchmark Comparison
DeepSeek V4 Flash 0731 has the higher score on 7 of 21 shared benchmarks; Muse Spark 1.2 leads on 14.
The largest observed score gap is 17.08 pts on Harvey's Legal Agent Benchmark , where Muse Spark 1.2 leads.
Reported ±1 standard-error ranges overlap on 8 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 | DeepSeek V4 Flash 0731 | Muse Spark 1.2 | Gap | Reported uncertainty |
|---|---|---|---|---|
| Vals Index | 47.96% ±1.24 | 49.29% ±1.10 | 1.32 pts | Reported ±1 SE ranges overlap |
| Harvey's Legal Agent Benchmark | 8.33% ±2.15 | 25.42% ±3.67 | 17.08 pts | Reported ±1 SE ranges do not overlap |
| Legal Research Bench | 30.29% ±3.19 | 43.75% ±3.45 | 13.46 pts | Reported ±1 SE ranges do not overlap |
| LegalBench | 77.71% ±0.50 | 85.26% ±0.45 | 7.55 pts | Reported ±1 SE ranges do not overlap |
| EMB | 56.98% ±3.39 | 56.98% ±3.13 | 0.00 pts | Reported ±1 SE ranges overlap |
| Finance Agent (v2) | 49.52% ±0.47 | 60.60% ±0.28 | 11.08 pts | Reported ±1 SE ranges do not overlap |
| Tax Agent Bench | 62.56% ±3.14 | 56.86% ±2.32 | 5.70 pts | Reported ±1 SE ranges do not overlap |
| TaxEval v2 | 70.69% ±0.90 | 80.38% ±0.76 | 9.69 pts | Reported ±1 SE ranges do not overlap |
| MedCode | 41.41% ±2.15 | 49.35% ±2.19 | 7.93 pts | Reported ±1 SE ranges do not overlap |
| MedScribe | 80.36% ±1.97 | 90.06% ±1.96 | 9.70 pts | Reported ±1 SE ranges do not overlap |
| ProofBench v1.1 | 56.00% ±4.99 | 43.00% ±4.98 | 13.00 pts | Reported ±1 SE ranges do not overlap |
| BioMysteryBench | 64.44% ±1.28 | 64.81% ±1.33 | 0.37 pts | Reported ±1 SE ranges overlap |
| MMLU Pro | 86.21% ±0.34 | 88.28% ±0.32 | 2.07 pts | Reported ±1 SE ranges do not overlap |
| Code Migration | 38.63% ±4.38 | 29.95% ±4.02 | 8.68 pts | Reported ±1 SE ranges do not overlap |
| IOI | 32.72% ±2.35 | 21.78% ±0.87 | 10.94 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 |
| SkillsBench | 50.67% ±4.43 | 53.04% ±4.42 | 2.37 pts | Reported ±1 SE ranges overlap |
| SWE-bench | 88.80% ±1.41 | 86.60% ±1.52 | 2.20 pts | Reported ±1 SE ranges overlap |
| Terminal-Bench 4.0 | 18.69% ±3.54 | 6.06% ±1.51 | 12.63 pts | Reported ±1 SE ranges do not overlap |
| Vibe Code Bench v1.1 | 74.74% ±3.44 | 79.10% ±3.31 | 4.36 pts | Reported ±1 SE ranges overlap |
| CyberBench v1.1 | 72.08% ±5.41 | 69.52% ±5.56 | 2.56 pts | Reported ±1 SE ranges overlap |
Performance by category
| Category | DeepSeek V4 Flash 0731 average | Muse Spark 1.2 average |
|---|---|---|
| Legal | 38.78% | 51.48% |
| Finance | 59.94% | 63.70% |
| Healthcare | 60.89% | 69.70% |
| Math | 56.00% | 43.00% |
| Science | 64.44% | 64.81% |
| Academic | 86.21% | 88.28% |
| Coding | 43.46% | 39.58% |
| Cyber | 72.08% | 69.52% |
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 | DeepSeek V4 Flash 0731 cost | Muse Spark 1.2 cost | DeepSeek V4 Flash 0731 latency | Muse Spark 1.2 latency |
|---|---|---|---|---|
| Vals Index | $0.85 | $1.94 | 43m06s | 22m56s |
| Harvey's Legal Agent Benchmark | $0.06 | $2.09 | 16m51s | 24m16s |
| Legal Research Bench | $0.28 | $0.51 | 21m07s | 5m07s |
| LegalBench | N/A | N/A | 3.95s | 30.10s |
| EMB | $0.38 | $2.31 | 26m14s | 17m07s |
| Finance Agent (v2) | $0.28 | $0.77 | 8m10s | 5m04s |
| Tax Agent Bench | $0.16 | $0.27 | 5m37s | 119.91s |
| TaxEval v2 | N/A | N/A | 72.20s | 48.31s |
| MedCode | N/A | N/A | 2m07s | 60.17s |
| MedScribe | N/A | N/A | 77.55s | 61.46s |
| ProofBench v1.1 | $0.04 | $0.44 | 15m34s | 7m49s |
| BioMysteryBench | $0.04 | $0.99 | 27m03s | 23m37s |
| MMLU Pro | N/A | N/A | 17.19s | 44.23s |
| Code Migration | $3.61 | $3.78 | 2h08m | 31m07s |
| IOI | $0.37 | $2.64 | 27m43s | 25m35s |
| ProgramBench | $1.29 | $2.47 | 5h39m | 26m38s |
| SkillsBench | $0.13 | $0.65 | 6m52s | 13m00s |
| SWE-bench | $0.01 | $0.55 | 3m22s | 8m07s |
| Terminal-Bench 4.0 | $0.77 | $4.14 | 1h46m | 1h18m |
| Vibe Code Bench v1.1 | $0.91 | $1.53 | 37m22s | 19m55s |
| CyberBench v1.1 | $0.19 | $2.15 | 15m28s | 20m21s |
Results available only for DeepSeek V4 Flash 0731
- Terminal-Bench Science
- GPQA Diamond
- LiveCodeBench
Results available only for Muse Spark 1.2
- MortgageTax
- MMMU Pro
- SAGE
- Public Benefits Bench v1.1