DeepSeek V4 Pro 0813 vs Grok 4.5: Benchmark Comparison
DeepSeek V4 Pro 0813 has the higher score on 12 of 22 shared benchmarks; Grok 4.5 leads on 9.
The largest observed score gap is 19.00 pts on ProofBench v1.1 , where DeepSeek V4 Pro 0813 leads.
Reported ±1 standard-error ranges overlap on 10 of 22 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 Pro 0813 | Grok 4.5 | Gap | Reported uncertainty |
|---|---|---|---|---|
| Vals Index | 47.63% ±1.12 | 44.72% ±1.17 | 2.91 pts | Reported ±1 SE ranges do not overlap |
| Harvey's Legal Agent Benchmark | 7.50% ±2.15 | 12.92% ±2.53 | 5.42 pts | Reported ±1 SE ranges do not overlap |
| Legal Research Bench | 40.87% ±3.42 | 37.98% ±3.37 | 2.88 pts | Reported ±1 SE ranges overlap |
| LegalBench | 82.36% ±0.44 | 85.97% ±0.41 | 3.60 pts | Reported ±1 SE ranges do not overlap |
| EMB | 52.80% ±3.06 | 52.91% ±3.11 | 0.11 pts | Reported ±1 SE ranges overlap |
| Finance Agent (v2) | 50.39% ±0.25 | 48.35% ±0.46 | 2.04 pts | Reported ±1 SE ranges do not overlap |
| Tax Agent Bench | 58.66% ±3.14 | 64.32% ±3.16 | 5.66 pts | Reported ±1 SE ranges overlap |
| TaxEval v2 | 73.06% ±0.87 | 71.67% ±0.89 | 1.39 pts | Reported ±1 SE ranges overlap |
| MedCode | 42.47% ±2.16 | 43.29% ±2.31 | 0.82 pts | Reported ±1 SE ranges overlap |
| MedScribe | 80.17% ±2.00 | 86.88% ±1.94 | 6.71 pts | Reported ±1 SE ranges do not overlap |
| ProofBench v1.1 | 50.00% ±5.03 | 31.00% ±4.65 | 19.00 pts | Reported ±1 SE ranges do not overlap |
| GPQA Diamond | 92.42% ±2.02 | 92.93% ±1.29 | 0.50 pts | Reported ±1 SE ranges overlap |
| MMLU Pro | 86.97% ±0.34 | 89.22% ±0.31 | 2.24 pts | Reported ±1 SE ranges do not overlap |
| Code Migration | 41.54% ±4.30 | 36.60% ±4.14 | 4.94 pts | Reported ±1 SE ranges overlap |
| IOI | 51.61% ±2.40 | 40.56% ±3.95 | 11.05 pts | Reported ±1 SE ranges do not overlap |
| LiveCodeBench | 87.53% ±0.96 | 87.35% ±0.96 | 0.17 pts | Reported ±1 SE ranges overlap |
| ProgramBench | 0.00% ±0.00 | 0.00% ±0.00 | 0.00 pts | Reported ±1 SE ranges overlap |
| SkillsBench | 53.83% ±4.49 | 66.03% ±4.60 | 12.21 pts | Reported ±1 SE ranges do not overlap |
| SWE-bench | 96.40% ±0.83 | 86.60% ±1.52 | 9.80 pts | Reported ±1 SE ranges do not overlap |
| Terminal-Bench 4.0 | 14.14% ±2.02 | 8.59% ±1.34 | 5.55 pts | Reported ±1 SE ranges do not overlap |
| Vibe Code Bench v1.1 | 82.30% ±3.15 | 69.00% ±4.48 | 13.30 pts | Reported ±1 SE ranges do not overlap |
| CyberBench v1.1 | 72.86% ±5.50 | 71.01% ±5.79 | 1.84 pts | Reported ±1 SE ranges overlap |
Performance by category
| Category | DeepSeek V4 Pro 0813 average | Grok 4.5 average |
|---|---|---|
| Legal | 43.58% | 45.62% |
| Finance | 58.73% | 59.31% |
| Healthcare | 61.32% | 65.09% |
| Math | 50.00% | 31.00% |
| Academic | 89.70% | 91.07% |
| Coding | 53.42% | 49.34% |
| Cyber | 72.86% | 71.01% |
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 Pro 0813 cost | Grok 4.5 cost | DeepSeek V4 Pro 0813 latency | Grok 4.5 latency |
|---|---|---|---|---|
| Vals Index | $3.35 | $2.46 | 1h04m | 15m44s |
| Harvey's Legal Agent Benchmark | $0.17 | $2.02 | 21m53s | 10m15s |
| Legal Research Bench | $1.09 | $1.09 | 42m07s | 15m54s |
| LegalBench | N/A | N/A | 20.05s | 67.88s |
| EMB | $1.08 | $1.55 | 37m27s | 18m28s |
| Finance Agent (v2) | $0.88 | $1.16 | 16m15s | 4m49s |
| Tax Agent Bench | $0.73 | $0.47 | 26m37s | 4m44s |
| TaxEval v2 | N/A | N/A | 3m46s | 3m15s |
| MedCode | N/A | N/A | 3m01s | 59.64s |
| MedScribe | N/A | N/A | 2m36s | 11m45s |
| ProofBench v1.1 | $0.06 | $0.57 | 12m12s | 11m09s |
| GPQA Diamond | N/A | N/A | 4m39s | 3m38s |
| MMLU Pro | N/A | N/A | 59.62s | 5m58s |
| Code Migration | $18.58 | $3.37 | 3h32m | 23m29s |
| IOI | $2.15 | $5.64 | 1h07m | 1h27m |
| LiveCodeBench | N/A | N/A | 4m33s | 9m49s |
| ProgramBench | $0.43 | $4.12 | 1h23m | 18m59s |
| SkillsBench | $0.37 | $0.58 | 11m15s | 2m16s |
| SWE-bench | $0.10 | $0.54 | 4m00s | 3m20s |
| Terminal-Bench 4.0 | $3.31 | $5.79 | 1h32m | 33m51s |
| Vibe Code Bench v1.1 | $0.36 | $4.10 | 1h06m | 13m50s |
| CyberBench v1.1 | $0.59 | $2.95 | 22m33s | 14m08s |
Results available only for DeepSeek V4 Pro 0813
- Terminal-Bench Science
- Vibe Code Bench 1-100
Results available only for Grok 4.5
- MortgageTax
- MMMU Pro
- SAGE
- SRE Bench
- Time Horizon Index: KSP