DeepSeek V4.1 Flash 1017 vs GPT-5.6 Sol: Benchmark Comparison
DeepSeek V4.1 Flash 1017 has the higher score on 4 of 9 shared benchmarks; GPT-5.6 Sol leads on 5.
The largest observed score gap is 13.26 pts on Tax Agent Bench , where DeepSeek V4.1 Flash 1017 leads.
Reported ±1 standard-error ranges overlap on 3 of 9 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 V4.1 Flash 1017 | GPT-5.6 Sol | Gap | Reported uncertainty |
|---|---|---|---|---|
| Vals Index | 56.58% ±1.27 | 58.01% ±1.02 | 1.42 pts | Reported ±1 SE ranges overlap |
| Harvey's Legal Agent Benchmark | 9.17% ±2.51 | 2.50% ±0.83 | 6.67 pts | Reported ±1 SE ranges do not overlap |
| Legal Research Bench | 45.19% ±3.46 | 48.08% ±3.47 | 2.88 pts | Reported ±1 SE ranges overlap |
| EMB | 61.80% ±3.19 | 72.34% ±2.37 | 10.53 pts | Reported ±1 SE ranges do not overlap |
| Finance Agent (v2) | 56.96% ±2.05 | 53.76% ±0.85 | 3.20 pts | Reported ±1 SE ranges do not overlap |
| Tax Agent Bench | 44.35% ±3.57 | 31.09% ±3.34 | 13.26 pts | Reported ±1 SE ranges do not overlap |
| Code Migration | 44.21% ±4.46 | 52.92% ±4.35 | 8.71 pts | Reported ±1 SE ranges overlap |
| Terminal-Bench 4.0 | 31.82% ±2.62 | 37.88% ±0.00 | 6.06 pts | Reported ±1 SE ranges do not overlap |
| Vibe Code Bench v1.1 | 87.73% ±2.75 | 80.50% ±3.72 | 7.23 pts | Reported ±1 SE ranges do not overlap |
Performance by category
| Category | DeepSeek V4.1 Flash 1017 average | GPT-5.6 Sol average |
|---|---|---|
| Legal | 27.18% | 25.29% |
| Finance | 54.37% | 52.39% |
| Coding | 54.59% | 57.10% |
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 V4.1 Flash 1017 cost | GPT-5.6 Sol cost | DeepSeek V4.1 Flash 1017 latency | GPT-5.6 Sol latency |
|---|---|---|---|---|
| Vals Index | $0.64 | $14.24 | 36m33s | 37m23s |
| Harvey's Legal Agent Benchmark | $0.54 | $10.37 | 21m39s | 22m50s |
| Legal Research Bench | $0.70 | $21.61 | 35m07s | 1h17m |
| EMB | $0.36 | $6.01 | 16m47s | 9m50s |
| Finance Agent (v2) | $0.46 | $1.25 | 14m39s | 19m25s |
| Tax Agent Bench | $0.46 | $7.51 | 21m04s | 44m41s |
| Code Migration | $1.86 | $24.54 | 2h11m | 59m12s |
| Terminal-Bench 4.0 | $0.54 | $7.98 | 41m10s | 33m46s |
| Vibe Code Bench v1.1 | $0.58 | $33.40 | 16m14s | 33m53s |
Results available only for DeepSeek V4.1 Flash 1017
None.
Results available only for GPT-5.6 Sol
- Vals RSI Index
- Vals Science Index
- LegalBench
- MortgageTax
- TaxEval v2
- MedCode
- MedScribe
- ProofBench v1.1
- BioMysteryBench
- MysteryMechanism
- Terminal-Bench-Science
- GPQA Diamond
- MMLU Pro
- MMMU Pro
- SAGE
- IOI
- LiveCodeBench
- ProgramBench
- SkillsBench
- SWE-bench
- Vibe Code Bench 1-100
- CyberBench v1.1
- SRE Bench
- Public Benefits Bench v1.1
- CUA-bench
- Time Horizon Index: KSP