# DeepSeek V4 Flash 0731 vs Hy4 preview

> Side-by-side benchmark comparison for DeepSeek V4 Flash 0731 and Hy4 preview across agentic, coding, multimodal, knowledge, reasoning, multilingual, and math tasks.

- Canonical page: https://benchlm.ai/compare/deepseek-v4-flash-0731-vs-hy4-preview
- Last updated: October 13, 2026

- Shared sourced benchmarks: 17
- HTML indexing: indexable
- Ranking lane: BenchAlign v5.8

## Quick Verdict

DeepSeek V4 Flash 0731 leads by point estimate. Conditional score ranges do not establish rank confidence. Choose using the category evidence and your own workload tests.

## Summary

- DeepSeek V4 Flash 0731 has the higher overall point estimate 62 to 60.86. Conditional score ranges do not establish rank confidence.
- The biggest single benchmark swing is Apex in Mathematics, with scores of 33.0% and 74.2%.
- Hy4 preview is the cheaper option on output tokens, which matters if you expect large responses or heavy interactive use.

## Model Snapshot

| Property | DeepSeek V4 Flash 0731 | Hy4 preview |
|----------|----------|----------|
| Creator | DeepSeek | Tencent |
| Type | Open Weight | Open Weight |
| Reasoning | Reasoning | Reasoning |
| Context | 1M | 1M |
| Overall Score | null | 60.86 |
| Benchmarks Covered | 41 | 30 |
| Pricing (input/output) | $0.14 / $0.28 | $0.00 / $0.00 |

## Category Breakdown

### Agentic

- Winner: Not comparable
- DeepSeek V4 Flash 0731 public-lane score: Coming soon
- Hy4 preview public-lane score: 54.5 (Estimated · #37/123)

| Benchmark | DeepSeek V4 Flash 0731 | Hy4 preview | Winner |
|-----------|-----------|-----------|--------|
| Terminal-Bench 2.0 | 56.9% | Coming soon | Coming soon |
| BrowseComp | 73.2% | Coming soon | Coming soon |
| HLE w/ tools | 45.1% | 55.4% | Hy4 preview |
| MCP Atlas | 69% | 83.7% | Hy4 preview |
| Toolathlon | 47.8% | Coming soon | Coming soon |
| Terminal-Bench 2.1 | 82.7% | 85.4% | Hy4 preview |
| CyberGym | 76.7% | 78.4% | Hy4 preview |
| Toolathlon-Verified | 70.3% | 74.1% | Hy4 preview |
| Agents' Last Exam | 25.2% | 22.8% | DeepSeek V4 Flash 0731 |
| AutomationBench | 25.1% | 32.1% | Hy4 preview |
| Terminal-Bench 2.1 (Vals) | 67.0% | 55.1% | DeepSeek V4 Flash 0731 |
| WideResearch | Coming soon | 83.9% | Coming soon |
| DRACO | Coming soon | 77.2% | Coming soon |
| APEX-Agents | Coming soon | 37.1% | Coming soon |
| skillsBench | Coming soon | 62.9% | Coming soon |
| JobBench | Coming soon | 61.7% | Coming soon |
| BankerToolBench | Coming soon | 78.6% | Coming soon |
| CWE-bench v1 | Coming soon | 53.0% | Coming soon |

### Coding

- Winner: Not comparable
- DeepSeek V4 Flash 0731 public-lane score: Coming soon
- Hy4 preview public-lane score: 57.1 (Supported · #26/148)

| Benchmark | DeepSeek V4 Flash 0731 | Hy4 preview | Winner |
|-----------|-----------|-----------|--------|
| LiveCodeBench Pass@1-COT | 91.6% | Coming soon | Coming soon |
| Codeforces | 3052.0 | Coming soon | Coming soon |
| SWE-bench Verified | 79% | Coming soon | Coming soon |
| SWE-bench Pro | 52.6% | 65.7% | Hy4 preview |
| SWE Multilingual | 73.3% | 82.9% | Hy4 preview |
| Terminal-Bench 2.0 | 56.9% | Coming soon | Coming soon |
| Terminal-Bench 2.1 | 82.7% | 85.4% | Hy4 preview |
| NL2Repo | 54.2% | 58.9% | Hy4 preview |
| DeepSWE | 54.4% | 64.3% | Hy4 preview |
| DSBench-FullStack | 68.7% | Coming soon | Coming soon |
| DSBench-Hard | 59.6% | Coming soon | Coming soon |
| VulcanBench v3 | 88.4% | Coming soon | Coming soon |
| OpenHarmony Bench | 53.8% | Coming soon | Coming soon |
| LiveCodeBench (Vals) | 87.3% | Coming soon | Coming soon |
| SWE-bench (Vals) | 88.8% | Coming soon | Coming soon |
| ProgramBench | Coming soon | 17.5% | Coming soon |
| PostTrain Bench | Coming soon | 35.6% | Coming soon |
| sweMarathon | Coming soon | 31.9% | Coming soon |

### Reasoning

- Winner: Not comparable
- DeepSeek V4 Flash 0731 public-lane score: 60.8 (Unranked · 4 rankable rows)
- Hy4 preview public-lane score: Coming soon

| Benchmark | DeepSeek V4 Flash 0731 | Hy4 preview | Winner |
|-----------|-----------|-----------|--------|
| MRCR 1M | 78.7% | Coming soon | Coming soon |
| CorpusQA 1M | 60.5% | Coming soon | Coming soon |
| ARC-AGI-1 | 89.00% | Coming soon | Coming soon |
| ARC-AGI-2 | 61.4% | Coming soon | Coming soon |
| CritPt | Coming soon | 16.9% | Coming soon |

### Multimodal & Grounded

- Winner: Not comparable
- DeepSeek V4 Flash 0731 public-lane score: Coming soon
- Hy4 preview public-lane score: 88.3 (Unranked · 1 rankable row)

| Benchmark | DeepSeek V4 Flash 0731 | Hy4 preview | Winner |
|-----------|-----------|-----------|--------|
| OfficeQA Pro | Coming soon | 66.2% | Coming soon |

### Knowledge

- Winner: Not comparable
- DeepSeek V4 Flash 0731 public-lane score: Coming soon
- Hy4 preview public-lane score: 63.6 (Estimated · #39/180)

| Benchmark | DeepSeek V4 Flash 0731 | Hy4 preview | Winner |
|-----------|-----------|-----------|--------|
| MMLU-Pro | 86.2% | Coming soon | Coming soon |
| Chinese-SimpleQA | 78.9% | Coming soon | Coming soon |
| GPQA | 88.1% | 92.3% | Hy4 preview |
| GPQA-D | 88.1% | 92.3% | Hy4 preview |
| HLE | 34.8% | 55.4% | Hy4 preview |
| GPQA Diamond (Vals) | 89.9% | Coming soon | Coming soon |
| MMLU-Pro (Vals) | 86.2% | Coming soon | Coming soon |
| HLE w/o tools | Coming soon | 43.4% | Coming soon |

### Mathematics

- Winner: Not comparable
- DeepSeek V4 Flash 0731 public-lane score: 87.1 (Unranked · 4 rankable rows)
- Hy4 preview public-lane score: Coming soon

| Benchmark | DeepSeek V4 Flash 0731 | Hy4 preview | Winner |
|-----------|-----------|-----------|--------|
| HMMT Feb 2026 | 94.8% | Coming soon | Coming soon |
| IMOAnswerBench | 88.4% | Coming soon | Coming soon |
| Apex | 33.0% | 74.2% | Hy4 preview |
| Apex Shortlist | 85.7% | Coming soon | Coming soon |

## FAQ

### Which is better overall, DeepSeek V4 Flash 0731 or Hy4 preview?

DeepSeek V4 Flash 0731 has the higher overall point estimate. Conditional score ranges do not establish rank confidence.

### How many benchmarks does DeepSeek V4 Flash 0731 cover on BenchLM?

DeepSeek V4 Flash 0731 currently has 41 sourced benchmark scores on BenchLM.

### How many benchmarks does Hy4 preview cover on BenchLM?

Hy4 preview currently has 30 sourced benchmark scores on BenchLM.

## Related Comparisons

- [DeepSeek V4 Flash 0731 vs Claude Opus 5.5](/compare/claude-opus-5-5-vs-deepseek-v4-flash-0731)
- [Hy4 preview vs Claude Opus 5.5](/compare/claude-opus-5-5-vs-hy4-preview)
- [DeepSeek V4 Flash 0731 vs GPT-6.1 Sol](/compare/deepseek-v4-flash-0731-vs-gpt-6-1-sol)
- [Hy4 preview vs GPT-6.1 Sol](/compare/gpt-6-1-sol-vs-hy4-preview)

## Explore More

- [DeepSeek V4 Flash 0731 profile](/models/deepseek-v4-flash-0731)
- [Hy4 preview profile](/models/hy4-preview)
- [Compare Pricing](/llm-pricing)
- [Alternative Finder](/tools/alternative-finder)
- [LLM Selector](/tools/llm-selector)
- [Overall Rankings](/best/overall)