Model comparison
DeepSeek V4 Pro Base vs Hy3 Preview
Head-to-head evidence from 0 shared benchmark results across 0 categories. Overall scores shown here use BenchLM's provisional ranking lane.
Evidence parity. DeepSeek V4 Pro Base and Hy3 Preview share 0 comparable benchmark results. 1 of 8 categories are comparable. 24 results are unique to DeepSeek V4 Pro Base; 22 to Hy3 Preview.
Updated July 14, 2026- Shared results
- 0
- DeepSeek V4 Pro Base only
- 24
- Hy3 Preview only
- 22
- Comparable categories
- 1 / 8
Pick Hy3 Preview if you want the stronger benchmark profile. DeepSeek V4 Pro Base only becomes the better choice if knowledge is the priority or you need the larger 1M context window.
Confidence note. This is a partial-evidence comparison with 0 shared benchmark results across 0 evidence categories; 1 of 8 categories currently have scoreable aggregates for both models. Treat the verdict as directional until coverage is more balanced.
Why this result
Hy3 Preview is clearly ahead on the provisional aggregate, 54 to 46. The gap is large enough that you do not need to squint at the spreadsheet to see the difference.
Hy3 Preview is the reasoning model in the pair, while DeepSeek V4 Pro Base is not. That usually helps on harder chain-of-thought-heavy tests, but it can also mean more latency and more token spend in real use. DeepSeek V4 Pro Base gives you the larger context window at 1M, compared with 256K for Hy3 Preview.
Operational comparison
Runtime and commercial metrics are compared only when both models have a complete sourced value.
| Metric | DeepSeek V4 Pro Base | Hy3 Preview | Comparison |
|---|---|---|---|
| Input / output priceUSD per 1M tokens | DeepSeek V4 Pro BaseNot available | Hy3 Preview$0 input / $0 output | A complete price comparison is not available. |
| Generation speedtokens per second | DeepSeek V4 Pro BaseNot available | Hy3 PreviewNot available | A complete speed comparison is not available. |
| First-answer latencyseconds to first token | DeepSeek V4 Pro BaseNot available | Hy3 PreviewNot available | A complete latency comparison is not available. |
| Context windowmaximum listed tokens | DeepSeek V4 Pro Base1M | Hy3 Preview256K | DeepSeek V4 Pro Base lists the larger context window. |
Benchmark Deep Dive
Agentic5 benchmarks
Coding7 benchmarks
| Benchmark | DeepSeek V4 Pro Base | Hy3 Preview | Result |
|---|---|---|---|
| BigCodeBenchSource | 59.2% | — | Not comparable |
| HumanEvalSource | 76.8% | — | Not comparable |
| SWE-bench VerifiedSource | — | 74.4% | Not comparable |
| Terminal-Bench 2.0Source | — | 54.4% | Not comparable |
| SciCodeSource | — | 41.2% | Not comparable |
| AA-SciCodeSource | — | 47.6% | Not comparable |
| AA Coding IndexSource | — | 58.8% | Not comparable |
Reasoning8 benchmarks
| Benchmark | DeepSeek V4 Pro Base | Hy3 Preview | Result |
|---|---|---|---|
| BBHSource | 87.5% | — | Not comparable |
| DROPSource | 88.7% | — | Not comparable |
| HellaSwagSource | 88.0% | — | Not comparable |
| WinoGrandeSource | 81.5% | — | Not comparable |
| CLUEWSCSource | 85.2% | — | Not comparable |
| LongBench v2Source | 51.5% | — | Not comparable |
| AA-LCRSource | — | 66.7% | Not comparable |
| CritPtSource | — | 4.9% | Not comparable |
KnowledgeDeepSeek V4 Pro Base wins21 benchmarks
| Benchmark | DeepSeek V4 Pro Base | Hy3 Preview | Result |
|---|---|---|---|
| AGIEvalSource | 83.1% | — | Not comparable |
| MMLUSource | 90.1% | — | Not comparable |
| MMLU-ReduxSource | 90.8% | — | Not comparable |
| MMLU-ProSource | 73.5% | — | Not comparable |
| MMMLUSource | 90.3% | — | Not comparable |
| C-EvalSource | 93.1% | — | Not comparable |
| CMMLUSource | 90.8% | — | Not comparable |
| MultiLoKoSource | 51.1% | — | Not comparable |
| SimpleQASource | 55.2% | — | Not comparable |
| SuperGPQASource | 53.9% | — | Not comparable |
| FACTS ParametricSource | 62.6% | — | Not comparable |
| TriviaQASource | 85.6% | — | Not comparable |
| Artificial Analysis Intelligence IndexSource | — | 41.2% | Not comparable |
| GPQASource | — | 87.2% | Not comparable |
| GPQA-DSource | — | 87.2% | Not comparable |
| HLESource | — | 25.5% | Not comparable |
| AA-Omniscience AccuracySource | — | 31.5% | Not comparable |
| AA-Omniscience Hallucination RateSource | — | 73.0% | Not comparable |
| AA-GPQA DiamondSource | — | 89.7% | Not comparable |
| AA-HLESource | — | 31.6% | Not comparable |
| AA-Omniscience IndexSource | — | -18.5% | Not comparable |
Math3 benchmarks
Multilingual1 benchmarks
| Benchmark | DeepSeek V4 Pro Base | Hy3 Preview | Result |
|---|---|---|---|
| MGSMSource | 84.4% | — | Not comparable |
Inst. Following1 benchmarks
| Benchmark | DeepSeek V4 Pro Base | Hy3 Preview | Result |
|---|---|---|---|
| IFBenchSource | — | 63.1% | Not comparable |
Frequently Asked Questions (2)
Which is better, DeepSeek V4 Pro Base or Hy3 Preview?
Hy3 Preview is ahead on BenchLM's provisional leaderboard, 54 to 46.
Which is better for knowledge tasks, DeepSeek V4 Pro Base or Hy3 Preview?
DeepSeek V4 Pro Base has the edge for knowledge tasks in this comparison, averaging 66.5 versus 33.8. Hy3 Preview stays close enough that the answer can still flip depending on your workload.
Related Comparisons
Explore More
The AI models change fast. We track them for you.
A weekly brief for engineers and researchers covering new models, ranking shifts, and pricing changes.
Free. No spam. Unsubscribe anytime.