DeepSeek V4 vs Mercury 2.5: Benchmark Comparison
DeepSeek V4 has the higher score on 9 of 10 shared benchmarks; Mercury 2.5 leads on 1.
The largest observed score gap is 46.25 pts on Vibe Code Bench v1.1 , where DeepSeek V4 leads.
Reported ±1 standard-error ranges overlap on 0 of 10 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 | Mercury 2.5 | Gap | Reported uncertainty |
|---|---|---|---|---|
| Legal Research Bench | 23.08% ±2.93 | 4.33% ±1.41 | 18.75 pts | Reported ±1 SE ranges do not overlap |
| LegalBench | 80.32% ±0.47 | 83.06% ±0.48 | 2.74 pts | Reported ±1 SE ranges do not overlap |
| Tax Agent Bench | 25.17% ±3.06 | 2.49% ±1.02 | 22.68 pts | Reported ±1 SE ranges do not overlap |
| MedCode | 40.45% ±2.12 | 31.33% ±1.95 | 9.13 pts | Reported ±1 SE ranges do not overlap |
| MedScribe | 75.14% ±2.00 | 55.09% ±2.10 | 20.05 pts | Reported ±1 SE ranges do not overlap |
| ProofBench v1.1 | 16.00% ±3.69 | 3.00% ±1.71 | 13.00 pts | Reported ±1 SE ranges do not overlap |
| SkillsBench | 51.27% ±4.60 | 18.06% ±3.08 | 33.21 pts | Reported ±1 SE ranges do not overlap |
| Terminal-Bench 4.0 | 11.11% ±1.34 | 0.00% ±0.00 | 11.11 pts | Reported ±1 SE ranges do not overlap |
| Vibe Code Bench v1.1 | 49.93% ±4.77 | 3.68% ±1.41 | 46.25 pts | Reported ±1 SE ranges do not overlap |
| Public Benefits Bench v1.1 | 62.92% ±1.26 | 45.26% ±1.29 | 17.66 pts | Reported ±1 SE ranges do not overlap |
Performance by category
| Category | DeepSeek V4 average | Mercury 2.5 average |
|---|---|---|
| Legal | 51.70% | 43.69% |
| Finance | 25.17% | 2.49% |
| Healthcare | 57.80% | 43.21% |
| Math | 16.00% | 3.00% |
| Coding | 37.44% | 7.25% |
| Social Mobility | 62.92% | 45.26% |
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 cost | Mercury 2.5 cost | DeepSeek V4 latency | Mercury 2.5 latency |
|---|---|---|---|---|
| Legal Research Bench | $0.68 | $0.06 | 17m05s | 47.61s |
| LegalBench | N/A | N/A | 3m04s | 3.91s |
| Tax Agent Bench | $0.72 | $0.02 | 20m45s | 29.67s |
| MedCode | N/A | N/A | 6m23s | 7.93s |
| MedScribe | N/A | N/A | 5m46s | 9.99s |
| ProofBench v1.1 | $0.03 | $0.04 | 7m03s | 3m22s |
| SkillsBench | $0.40 | $0.10 | 8m16s | 49.36s |
| Terminal-Bench 4.0 | $3.45 | $0.37 | 1h26m | 15m14s |
| Vibe Code Bench v1.1 | $2.21 | $0.38 | 56m59s | 3m13s |
| Public Benefits Bench v1.1 | $0.46 | $0.07 | 15m22s | 68.33s |
Results available only for DeepSeek V4
- TaxEval v2
- GPQA Diamond
- MMLU Pro
- LiveCodeBench
- ProgramBench
- SWE-bench
Results available only for Mercury 2.5
- Terminal-Bench-Science
- IOI
- CyberBench v1.1