# Claude Haiku 5.5 vs DeepSeek V4 Pro

> Side-by-side benchmark comparison for Claude Haiku 5.5 and DeepSeek V4 Pro across agentic, coding, multimodal, knowledge, reasoning, multilingual, and math tasks.

- Canonical page: https://benchlm.ai/compare/claude-haiku-5-5-vs-deepseek-v4-pro
- Last updated: November 21, 2026

- Shared sourced benchmarks: 7
- HTML indexing: indexable
- Ranking lane: BenchAlign v6.0

## Quick Verdict

Pick Claude Haiku 5.5 if you want the stronger benchmark profile. DeepSeek V4 Pro only makes more sense when its price, context window, or workload-specific category wins matter more than the overall score.

## Summary

- Claude Haiku 5.5 leads overall 63.5 to 61.6.
- The clearest category separation is in knowledge, where the averages are 50.2 for Claude Haiku 5.5 and 59.7 for DeepSeek V4 Pro.
- The biggest single benchmark swing is Terminal-Bench 4.1 in Agentic, with scores of 40.00% and 6.70%.
- Claude Haiku 5.5 is the cheaper option on output tokens, which matters if you expect large responses or heavy interactive use.

## Model Snapshot

| Property | Claude Haiku 5.5 | DeepSeek V4 Pro |
|----------|----------|----------|
| Creator | Anthropic | DeepSeek |
| Type | Proprietary | Open Weight |
| Reasoning | Reasoning | Reasoning |
| Context | 1M | 1M |
| Overall Score | 63.5 | 61.6 |
| Benchmarks Covered | 27 | 46 |
| Pricing (input/output) | $0.10 / $0.50 | $1.32 / $3.96 |

## Category Breakdown

### Agentic

- Winner: Claude Haiku 5.5
- Claude Haiku 5.5 public-lane score: 68.3 (Supported · #25/182)
- DeepSeek V4 Pro public-lane score: 60 (Supported · #55/182)

| Benchmark | Claude Haiku 5.5 | DeepSeek V4 Pro | Winner |
|-----------|-----------|-----------|--------|
| Gert Labs | 58.62% | 52.74% | Claude Haiku 5.5 |
| Terminal-Bench 4.1 | 40.00% | 6.70% | Claude Haiku 5.5 |
| Agents' Last Exam (Linux) | 57.0% | Coming soon | Coming soon |
| Terminal-Bench 4.0 | 39.20% | Coming soon | Coming soon |
| HLE w/ tools | 57.4% | 60.0% | DeepSeek V4 Pro |
| MazeBench | 1% | Coming soon | Coming soon |
| Hermes Index | 37.0% | Coming soon | Coming soon |
| Hermes Bench | 71.0% | Coming soon | Coming soon |
| OSWorld 2.1 Offline | 72.4% | Coming soon | Coming soon |
| Finch | Coming soon | 67.4% | Coming soon |
| Terminal-Bench 2.0 | Coming soon | 67.9% | Coming soon |
| Terminal-Bench 2.1 | Coming soon | 87.9% | Coming soon |
| BrowseComp | Coming soon | 83.4% | Coming soon |
| MCP Atlas | Coming soon | 73.6% | Coming soon |
| Toolathlon | Coming soon | 51.8% | Coming soon |
| CyberGym | Coming soon | 83.3% | Coming soon |
| Toolathlon-Verified | Coming soon | 74.1% | Coming soon |
| Agents' Last Exam | Coming soon | 25.7% | Coming soon |
| AutomationBench | Coming soon | 31.8% | Coming soon |
| Terminal-Bench 2.1 (Vals) | Coming soon | 54.7% | Coming soon |

### Coding

- Winner: Claude Haiku 5.5
- Claude Haiku 5.5 public-lane score: 67.9 (Supported · #28/180)
- DeepSeek V4 Pro public-lane score: 59.6 (Supported · #56/180)

| Benchmark | Claude Haiku 5.5 | DeepSeek V4 Pro | Winner |
|-----------|-----------|-----------|--------|
| FrontierCode 1.1 Main | 46.4% | Coming soon | Coming soon |
| Bug Hunt Bench | 21.5 fixes | Coming soon | Coming soon |
| FrontierSWE v2 | 43.8% | Coming soon | Coming soon |
| FrontierCode 2.0 | 28.9% | Coming soon | Coming soon |
| CursorBench 4.0 | 48.4% | Coming soon | Coming soon |
| VulcanBench Frontier v4 | 92.2% | Coming soon | Coming soon |
| CursorBench 4.1 | 48.4% | Coming soon | Coming soon |
| LiveCodeBench Pass@1-COT | Coming soon | 93.5% | Coming soon |
| Codeforces | Coming soon | 3206.0 | Coming soon |
| SWE-bench Verified | Coming soon | 80.6% | Coming soon |
| SWE-bench Pro | Coming soon | 55.4% | Coming soon |
| SWE Multilingual | Coming soon | 76.2% | Coming soon |
| Terminal-Bench 2.0 | Coming soon | 67.9% | Coming soon |
| Vibe Code Bench | Coming soon | 49.93% | Coming soon |
| Terminal-Bench 2.1 | Coming soon | 87.9% | Coming soon |
| NL2Repo | Coming soon | 61.5% | Coming soon |
| DeepSWE | Coming soon | 62.7% | Coming soon |
| DSBench-FullStack | Coming soon | 71.1% | Coming soon |
| DSBench-Hard | Coming soon | 67.2% | Coming soon |
| OpenHarmony Bench v1.0 | Coming soon | 59.0% | Coming soon |
| LiveCodeBench (Vals) | Coming soon | 87.5% | Coming soon |
| SWE-bench (Vals) | Coming soon | 96.4% | Coming soon |

### Reasoning

- Winner: DeepSeek V4 Pro
- Claude Haiku 5.5 public-lane score: 62.6 (Supported · #38/161)
- DeepSeek V4 Pro public-lane score: 64.6 (Supported · #32/161)

| Benchmark | Claude Haiku 5.5 | DeepSeek V4 Pro | Winner |
|-----------|-----------|-----------|--------|
| ARC-AGI-1 | 89.00% | 90.00% | DeepSeek V4 Pro |
| ARC-AGI-2 | 61.3% | 61.3% | Tie |
| ARC-AGI-3 | 0.5% | Coming soon | Coming soon |
| ARC-AGI-3 Provider Adapter | 2.82% | Coming soon | Coming soon |
| MRCR 1M | Coming soon | 83.5% | Coming soon |
| CorpusQA 1M | Coming soon | 62.0% | Coming soon |

### Multimodal & Grounded

- Winner: Directional only
- Claude Haiku 5.5 public-lane score: 63.9 (Supported · #39/119)
- DeepSeek V4 Pro public-lane score: 60.7 (Estimated · #54/119)

| Benchmark | Claude Haiku 5.5 | DeepSeek V4 Pro | Winner |
|-----------|-----------|-----------|--------|
| ParseBench-2 | 71.24% | Coming soon | Coming soon |
| PerceptionBench (tools) | 78.6% | Coming soon | Coming soon |
| ParseBench-1 | 56.06% | Coming soon | Coming soon |
| Chartography (no tools) | 46.4% | Coming soon | Coming soon |

### Knowledge

- Winner: DeepSeek V4 Pro
- Claude Haiku 5.5 public-lane score: 50.2 (Supported · #84/158)
- DeepSeek V4 Pro public-lane score: 59.7 (Supported · #46/158)

| Benchmark | Claude Haiku 5.5 | DeepSeek V4 Pro | Winner |
|-----------|-----------|-----------|--------|
| HLE w/o tools | 45.9% | Coming soon | Coming soon |
| MMLU-Pro | Coming soon | 87.5% | Coming soon |
| Chinese-SimpleQA | Coming soon | 84.4% | Coming soon |
| GPQA | Coming soon | 90.1% | Coming soon |
| GPQA-D | Coming soon | 90.1% | Coming soon |
| HLE | Coming soon | 42.7% | Coming soon |
| GPQA Diamond (Vals) | Coming soon | 92.4% | Coming soon |
| MMLU-Pro (Vals) | Coming soon | 87.0% | Coming soon |

### Multilingual

- Winner: Directional only
- Claude Haiku 5.5 public-lane score: 56.1 (Estimated · #36/88)
- DeepSeek V4 Pro public-lane score: 57.3 (Estimated · #31/88)

### Instruction Following

- Winner: Claude Haiku 5.5
- Claude Haiku 5.5 public-lane score: 58.3 (Supported · #37/143)
- DeepSeek V4 Pro public-lane score: 56.6 (Supported · #47/143)

### Mathematics

- Winner: Claude Haiku 5.5
- Claude Haiku 5.5 public-lane score: 62.1 (Supported · #29/125)
- DeepSeek V4 Pro public-lane score: 59.4 (Supported · #34/125)

| Benchmark | Claude Haiku 5.5 | DeepSeek V4 Pro | Winner |
|-----------|-----------|-----------|--------|
| FrontierMath v2 (Tiers 1-3) | 75.088% | 64.561% | Claude Haiku 5.5 |
| FrontierMath v2 (Tier 4) | 46.341% | 26.829% | Claude Haiku 5.5 |
| HMMT Feb 2026 | Coming soon | 95.2% | Coming soon |
| IMOAnswerBench | Coming soon | 89.8% | Coming soon |
| Apex | Coming soon | 38.3% | Coming soon |
| Apex Shortlist | Coming soon | 90.2% | Coming soon |

## FAQ

### Which is better overall, Claude Haiku 5.5 or DeepSeek V4 Pro?

Claude Haiku 5.5 is ahead overall on BenchLM right now.

### Where is the biggest gap between Claude Haiku 5.5 and DeepSeek V4 Pro?

The widest category gap is in knowledge, where the averages are 50.2 for Claude Haiku 5.5 and 59.7 for DeepSeek V4 Pro.

### How many benchmarks does Claude Haiku 5.5 cover on BenchLM?

Claude Haiku 5.5 currently has 27 sourced benchmark scores on BenchLM.

### How many benchmarks does DeepSeek V4 Pro cover on BenchLM?

DeepSeek V4 Pro currently has 46 sourced benchmark scores on BenchLM.

## Related Comparisons

- [Claude Haiku 5.5 vs Claude Fable 5.2](/compare/claude-fable-5-2-vs-claude-haiku-5-5)
- [DeepSeek V4 Pro vs Claude Fable 5.2](/compare/claude-fable-5-2-vs-deepseek-v4-pro)
- [Claude Haiku 5.5 vs Gemini 4 Neon](/compare/claude-haiku-5-5-vs-gemini-4-neon)
- [DeepSeek V4 Pro vs Gemini 4 Neon](/compare/deepseek-v4-pro-vs-gemini-4-neon)

## Explore More

- [Claude Haiku 5.5 profile](/models/claude-haiku-5-5)
- [DeepSeek V4 Pro profile](/models/deepseek-v4-pro)
- [Compare Pricing](/llm-pricing)
- [Alternative Finder](/tools/alternative-finder)
- [LLM Selector](/tools/llm-selector)
- [Overall Rankings](/best/overall)
