# DeepSeek V4 Flash 0731 vs Step 3.7 Flash

> Side-by-side benchmark comparison for DeepSeek V4 Flash 0731 and Step 3.7 Flash across agentic, coding, multimodal, knowledge, reasoning, multilingual, and math tasks.

- Canonical page: https://benchlm.ai/compare/deepseek-v4-flash-0731-vs-step-3-7-flash
- Last updated: September 27, 2026

- Shared sourced benchmarks: 6
- HTML indexing: indexable
- Ranking lane: BenchAlign v5.7

## Quick Verdict

Pick DeepSeek V4 Flash 0731 if you want the stronger benchmark profile. Step 3.7 Flash only makes more sense when its price, context window, or workload-specific category wins matter more than the overall score.

## Summary

- DeepSeek V4 Flash 0731 leads overall 57 to 41.41.
- The clearest category separation is in reasoning, where the averages are 57 for DeepSeek V4 Flash 0731 and 72.7 for Step 3.7 Flash.
- The biggest single benchmark swing is Terminal-Bench 2.1 in Agentic, with scores of 82.7% and 59.5%.
- DeepSeek V4 Flash 0731 is the cheaper option on output tokens, which matters if you expect large responses or heavy interactive use.
- DeepSeek V4 Flash 0731 also has the larger context window at 1M.

## Model Snapshot

| Property | DeepSeek V4 Flash 0731 | Step 3.7 Flash |
|----------|----------|----------|
| Creator | DeepSeek | StepFun |
| Type | Open Weight | Open Weight |
| Reasoning | Reasoning | Reasoning |
| Context | 1M | 256K |
| Overall Score | null | 41.41 |
| Benchmarks Covered | 42 | 11 |
| Pricing (input/output) | $0.14 / $0.28 | $0.20 / $1.15 |

## Category Breakdown

### Agentic

- Winner: Not comparable
- DeepSeek V4 Flash 0731 public-lane score: Coming soon
- Step 3.7 Flash public-lane score: 35.2 (Estimated · #56/105)

| Benchmark | DeepSeek V4 Flash 0731 | Step 3.7 Flash | Winner |
|-----------|-----------|-----------|--------|
| Terminal-Bench 2.0 | 56.9% | Coming soon | Coming soon |
| BrowseComp | 73.2% | 75.8% | Step 3.7 Flash |
| HLE w/ tools | 45.1% | 47.2% | Step 3.7 Flash |
| MCP Atlas | 69% | Coming soon | Coming soon |
| Toolathlon | 47.8% | 49.5% | Step 3.7 Flash |
| Terminal-Bench 2.1 | 82.7% | 59.5% | DeepSeek V4 Flash 0731 |
| CyberGym | 76.7% | Coming soon | Coming soon |
| Toolathlon-Verified | 70.3% | Coming soon | Coming soon |
| Agents' Last Exam | 25.2% | Coming soon | Coming soon |
| AutomationBench | 25.1% | Coming soon | Coming soon |
| Terminal-Bench 2.1 (Vals) | 67.0% | Coming soon | Coming soon |
| DeepSearchQA | Coming soon | 92.8% | Coming soon |
| Claw-Eval | Coming soon | 67.1% | Coming soon |
| Gert Labs | Coming soon | 51.57% | Coming soon |

### Coding

- Winner: Not comparable
- DeepSeek V4 Flash 0731 public-lane score: Coming soon
- Step 3.7 Flash public-lane score: 32.4 (Estimated · #81/135)

| Benchmark | DeepSeek V4 Flash 0731 | Step 3.7 Flash | 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% | 56.3% | Step 3.7 Flash |
| SWE Multilingual | 73.3% | Coming soon | Coming soon |
| Terminal-Bench 2.0 | 56.9% | Coming soon | Coming soon |
| Terminal-Bench 2.1 | 82.7% | 59.5% | DeepSeek V4 Flash 0731 |
| NL2Repo | 54.2% | Coming soon | Coming soon |
| DeepSWE | 54.4% | Coming soon | Coming soon |
| 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 |

### Reasoning

- Winner: Not comparable
- DeepSeek V4 Flash 0731 public-lane score: 57 (Unranked · 4 rankable rows)
- Step 3.7 Flash public-lane score: 72.7 (Unranked · 2 rankable rows)

| Benchmark | DeepSeek V4 Flash 0731 | Step 3.7 Flash | 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 |

### Multimodal & Grounded

- Winner: Not comparable
- DeepSeek V4 Flash 0731 public-lane score: Coming soon
- Step 3.7 Flash public-lane score: 71 (Unranked · 3 rankable rows)

| Benchmark | DeepSeek V4 Flash 0731 | Step 3.7 Flash | Winner |
|-----------|-----------|-----------|--------|
| SimpleVQA | Coming soon | 79.2% | Coming soon |
| V* | Coming soon | 95.3% | Coming soon |

### Knowledge

- Winner: Not comparable
- DeepSeek V4 Flash 0731 public-lane score: Coming soon
- Step 3.7 Flash public-lane score: 43.1 (Estimated · #78/158)

| Benchmark | DeepSeek V4 Flash 0731 | Step 3.7 Flash | Winner |
|-----------|-----------|-----------|--------|
| MMLU-Pro | 86.2% | Coming soon | Coming soon |
| SimpleQA | 34.1% | Coming soon | Coming soon |
| Chinese-SimpleQA | 78.9% | Coming soon | Coming soon |
| GPQA | 88.1% | Coming soon | Coming soon |
| GPQA-D | 88.1% | Coming soon | Coming soon |
| HLE | 34.8% | Coming soon | Coming soon |
| GPQA Diamond (Vals) | 89.9% | Coming soon | Coming soon |
| MMLU-Pro (Vals) | 86.2% | Coming soon | Coming soon |

### Instruction Following

- Winner: Not comparable
- DeepSeek V4 Flash 0731 public-lane score: Coming soon
- Step 3.7 Flash public-lane score: 80.6 (#53/124)

### Mathematics

- Winner: Not comparable
- DeepSeek V4 Flash 0731 public-lane score: 79.9 (Unranked · 4 rankable rows)
- Step 3.7 Flash public-lane score: Coming soon

| Benchmark | DeepSeek V4 Flash 0731 | Step 3.7 Flash | Winner |
|-----------|-----------|-----------|--------|
| HMMT Feb 2026 | 94.8% | Coming soon | Coming soon |
| IMOAnswerBench | 88.4% | Coming soon | Coming soon |
| Apex | 33.0% | Coming soon | Coming soon |
| Apex Shortlist | 85.7% | Coming soon | Coming soon |

## FAQ

### Which is better overall, DeepSeek V4 Flash 0731 or Step 3.7 Flash?

DeepSeek V4 Flash 0731 is ahead overall on BenchLM right now.

### Where is the biggest gap between DeepSeek V4 Flash 0731 and Step 3.7 Flash?

The widest category gap is in reasoning, where the averages are 57 for DeepSeek V4 Flash 0731 and 72.7 for Step 3.7 Flash.

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

DeepSeek V4 Flash 0731 currently has 42 sourced benchmark scores on BenchLM.

### How many benchmarks does Step 3.7 Flash cover on BenchLM?

Step 3.7 Flash currently has 11 sourced benchmark scores on BenchLM.

## Related Comparisons

- [DeepSeek V4 Flash 0731 vs GPT-6 Astra](/compare/deepseek-v4-flash-0731-vs-gpt-6-astra)
- [Step 3.7 Flash vs GPT-6 Astra](/compare/gpt-6-astra-vs-step-3-7-flash)
- [DeepSeek V4 Flash 0731 vs Claude Opus 5.5](/compare/claude-opus-5-5-vs-deepseek-v4-flash-0731)
- [Step 3.7 Flash vs Claude Opus 5.5](/compare/claude-opus-5-5-vs-step-3-7-flash)

## Explore More

- [DeepSeek V4 Flash 0731 profile](/models/deepseek-v4-flash-0731)
- [Step 3.7 Flash profile](/models/step-3-7-flash)
- [Compare Pricing](/llm-pricing)
- [Alternative Finder](/tools/alternative-finder)
- [LLM Selector](/tools/llm-selector)
- [Overall Rankings](/best/overall)