# Claude Opus 4.5 vs DeepSeek V4 Pro 0813

> Side-by-side benchmark comparison for Claude Opus 4.5 and DeepSeek V4 Pro 0813 across agentic, coding, multimodal, knowledge, reasoning, multilingual, and math tasks.

- Canonical page: https://benchlm.ai/compare/claude-opus-4-5-vs-deepseek-v4-pro-0813
- Last updated: September 27, 2026

- Shared sourced benchmarks: 12
- HTML indexing: indexable
- Ranking lane: BenchAlign v5.7

## Quick Verdict

Pick DeepSeek V4 Pro 0813 if you want the stronger benchmark profile. Claude Opus 4.5 only makes more sense when its price, context window, or workload-specific category wins matter more than the overall score.

## Summary

- DeepSeek V4 Pro 0813 leads overall 63.48 to 55.09.
- The clearest category separation is in agentic, where the averages are 30.9 for Claude Opus 4.5 and 55 for DeepSeek V4 Pro 0813.
- The biggest single benchmark swing is CyberGym in Agentic, with scores of 50.6% and 83.3%.
- DeepSeek V4 Pro 0813 is the cheaper option on output tokens, which matters if you expect large responses or heavy interactive use.
- DeepSeek V4 Pro 0813 also has the larger context window at 1M.

## Model Snapshot

| Property | Claude Opus 4.5 | DeepSeek V4 Pro 0813 |
|----------|----------|----------|
| Creator | Anthropic | DeepSeek |
| Type | Proprietary | Open Weight |
| Reasoning | Non-Reasoning | Reasoning |
| Context | 200K | 1M |
| Overall Score | 55.09 | 63.48 |
| Benchmarks Covered | 45 | 42 |
| Pricing (input/output) | $5.00 / $25.00 | $1.32 / $3.96 |

## Category Breakdown

### Agentic

- Winner: Directional only
- Claude Opus 4.5 public-lane score: 30.9 (Estimated · #65/105)
- DeepSeek V4 Pro 0813 public-lane score: 55 (Supported · #29/105)

| Benchmark | Claude Opus 4.5 | DeepSeek V4 Pro 0813 | Winner |
|-----------|-----------|-----------|--------|
| Terminal-Bench 2.0 | 59.3% | 67.9% | DeepSeek V4 Pro 0813 |
| OSWorld-Verified | 66.3% | Coming soon | Coming soon |
| OSWorld | 66.3% | Coming soon | Coming soon |
| Claw-Eval | 59.6% | Coming soon | Coming soon |
| QwenClawBench | 52.3% | Coming soon | Coming soon |
| τ³-bench results | 70.2% | Coming soon | Coming soon |
| VITA-Bench | 23.3% | Coming soon | Coming soon |
| DeepPlanning | 26.4% | Coming soon | Coming soon |
| Toolathlon | 43.5% | 51.8% | DeepSeek V4 Pro 0813 |
| MCP Atlas | 42.3% | 73.6% | DeepSeek V4 Pro 0813 |
| MCP-Tasks | 71.8% | Coming soon | Coming soon |
| WideResearch | 76.4% | Coming soon | Coming soon |
| CyberGym | 50.6% | 83.3% | DeepSeek V4 Pro 0813 |
| Gert Labs | 64.23% | Coming soon | Coming soon |
| JobBench | 32.3% | Coming soon | Coming soon |
| Terminal-Bench 2.1 | Coming soon | 87.9% | Coming soon |
| BrowseComp | Coming soon | 83.4% | Coming soon |
| HLE w/ tools | Coming soon | 60.0% | 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: Directional only
- Claude Opus 4.5 public-lane score: 43.5 (Estimated · #50/135)
- DeepSeek V4 Pro 0813 public-lane score: 50.3 (Supported · #39/135)

| Benchmark | Claude Opus 4.5 | DeepSeek V4 Pro 0813 | Winner |
|-----------|-----------|-----------|--------|
| SWE-bench Verified | 80.9% | 80.6% | Claude Opus 4.5 |
| LiveCodeBench v6 | 84.8% | Coming soon | Coming soon |
| SWE-bench Pro | 57.1% | 55.4% | Claude Opus 4.5 |
| SWE Multilingual | 77.5% | 76.2% | Claude Opus 4.5 |
| NL2Repo | 43.2% | 61.5% | DeepSeek V4 Pro 0813 |
| LiveCodeBench Pass@1-COT | Coming soon | 93.5% | Coming soon |
| Codeforces | Coming soon | 3206.0 | 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 |
| 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 | 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: Not comparable
- Claude Opus 4.5 public-lane score: 73.2 (Unranked · 4 rankable rows)
- DeepSeek V4 Pro 0813 public-lane score: 56.9 (Unranked · 4 rankable rows)

| Benchmark | Claude Opus 4.5 | DeepSeek V4 Pro 0813 | Winner |
|-----------|-----------|-----------|--------|
| LongBench v2 | 64.4% | Coming soon | Coming soon |
| AI-Needle | 74% | Coming soon | Coming soon |
| MRCR 1M | Coming soon | 83.5% | Coming soon |
| CorpusQA 1M | Coming soon | 62.0% | Coming soon |
| ARC-AGI-1 | Coming soon | 90.00% | Coming soon |
| ARC-AGI-2 | Coming soon | 61.3% | Coming soon |

### Multimodal & Grounded

- Winner: Not comparable
- Claude Opus 4.5 public-lane score: 23.5 (#48/50)
- DeepSeek V4 Pro 0813 public-lane score: Coming soon

| Benchmark | Claude Opus 4.5 | DeepSeek V4 Pro 0813 | Winner |
|-----------|-----------|-----------|--------|
| MMMU-Pro | 70.6% | Coming soon | Coming soon |
| MathVision | 74.3% | Coming soon | Coming soon |
| CharXiv | 68.5% | Coming soon | Coming soon |
| VideoMMMU | 84.4% | Coming soon | Coming soon |
| ScreenSpot Pro | 45.7% | Coming soon | Coming soon |
| V* | 67.0% | Coming soon | Coming soon |

### Knowledge

- Winner: Directional only
- Claude Opus 4.5 public-lane score: 49.4 (Estimated · #58/158)
- DeepSeek V4 Pro 0813 public-lane score: 63.4 (Estimated · #29/158)

| Benchmark | Claude Opus 4.5 | DeepSeek V4 Pro 0813 | Winner |
|-----------|-----------|-----------|--------|
| GPQA | 87% | 90.1% | DeepSeek V4 Pro 0813 |
| SuperGPQA | 70.6% | Coming soon | Coming soon |
| MMLU-Pro | 89.5% | 87.5% | Claude Opus 4.5 |
| MMLU-Redux | 96.6% | Coming soon | Coming soon |
| C-Eval | 92.2% | Coming soon | Coming soon |
| HLE | 30.8% | 42.7% | DeepSeek V4 Pro 0813 |
| SimpleQA | Coming soon | 57.9% | Coming soon |
| Chinese-SimpleQA | Coming soon | 84.4% | Coming soon |
| GPQA-D | Coming soon | 90.1% | Coming soon |
| GPQA Diamond (Vals) | Coming soon | 92.4% | Coming soon |
| MMLU-Pro (Vals) | Coming soon | 87.0% | Coming soon |

### Multilingual

- Winner: Not comparable
- Claude Opus 4.5 public-lane score: 82.9 (#2/12)
- DeepSeek V4 Pro 0813 public-lane score: Coming soon

| Benchmark | Claude Opus 4.5 | DeepSeek V4 Pro 0813 | Winner |
|-----------|-----------|-----------|--------|
| MMLU-ProX | 85.7% | Coming soon | Coming soon |
| NOVA-63 | 56.7% | Coming soon | Coming soon |

### Instruction Following

- Winner: Not comparable
- Claude Opus 4.5 public-lane score: 30.6 (#114/124)
- DeepSeek V4 Pro 0813 public-lane score: Coming soon

| Benchmark | Claude Opus 4.5 | DeepSeek V4 Pro 0813 | Winner |
|-----------|-----------|-----------|--------|
| IFEval | 90.9% | Coming soon | Coming soon |
| IFBench | 58% | Coming soon | Coming soon |

### Mathematics

- Winner: Not comparable
- Claude Opus 4.5 public-lane score: 57.9 (#6/7)
- DeepSeek V4 Pro 0813 public-lane score: 80.2 (Unranked · 4 rankable rows)

| Benchmark | Claude Opus 4.5 | DeepSeek V4 Pro 0813 | Winner |
|-----------|-----------|-----------|--------|
| AIME26 | 95.1% | Coming soon | Coming soon |
| HMMT Feb 2025 | 92.9% | Coming soon | Coming soon |
| HMMT Nov 2025 | 93.3% | Coming soon | Coming soon |
| HMMT Feb 2026 | 85.3% | 95.2% | DeepSeek V4 Pro 0813 |
| MMAnswerBench | 84.0% | Coming soon | Coming soon |
| FrontierMath v2 (Tiers 1-3) | 20.690% | Coming soon | Coming soon |
| FrontierMath v2 (Tier 4) | 4.167% | Coming soon | 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 Opus 4.5 or DeepSeek V4 Pro 0813?

DeepSeek V4 Pro 0813 is ahead overall on BenchLM right now.

### Where is the biggest gap between Claude Opus 4.5 and DeepSeek V4 Pro 0813?

The widest category gap is in agentic, where the averages are 30.9 for Claude Opus 4.5 and 55 for DeepSeek V4 Pro 0813.

### How many benchmarks does Claude Opus 4.5 cover on BenchLM?

Claude Opus 4.5 currently has 45 sourced benchmark scores on BenchLM.

### How many benchmarks does DeepSeek V4 Pro 0813 cover on BenchLM?

DeepSeek V4 Pro 0813 currently has 42 sourced benchmark scores on BenchLM.

## Related Comparisons

- [Claude Opus 4.5 vs Claude Opus 4.5 Thinking](/compare/claude-opus-4-5-vs-claude-opus-4-5-thinking)
- [DeepSeek V4 Pro 0813 vs Claude Opus 4.5 Thinking](/compare/claude-opus-4-5-thinking-vs-deepseek-v4-pro-0813)
- [Claude Opus 4.5 vs GPT-6 Astra](/compare/claude-opus-4-5-vs-gpt-6-astra)
- [DeepSeek V4 Pro 0813 vs GPT-6 Astra](/compare/deepseek-v4-pro-0813-vs-gpt-6-astra)

## Explore More

- [Claude Opus 4.5 profile](/models/claude-opus-4-5)
- [DeepSeek V4 Pro 0813 profile](/models/deepseek-v4-pro-0813)
- [Compare Pricing](/llm-pricing)
- [Alternative Finder](/tools/alternative-finder)
- [LLM Selector](/tools/llm-selector)
- [Overall Rankings](/best/overall)