# DeepSeek-R1 vs DeepSeek-V2.5

> Side-by-side benchmark comparison for DeepSeek-R1 and DeepSeek-V2.5 across agentic, coding, multimodal, knowledge, reasoning, multilingual, and math tasks.

- Canonical page: https://benchlm.ai/compare/deepseek-r1-vs-deepseek-v2-5
- Last updated: October 14, 2026

- Shared sourced benchmarks: 0
- HTML indexing: indexable
- Ranking lane: BenchAlign v5.8

## Quick Verdict

No benchmark verdict yet: there is no shared sourced benchmark row for DeepSeek-R1 and DeepSeek-V2.5.

## Summary

- DeepSeek-R1 and DeepSeek-V2.5 do not currently share a sourced benchmark row. The mirror therefore compares metadata, pricing, context windows, and each model's separately reported coverage without naming a benchmark winner.

## Model Snapshot

| Property | DeepSeek-R1 | DeepSeek-V2.5 |
|----------|----------|----------|
| Creator | DeepSeek | DeepSeek |
| Type | Open Weight | Open Weight |
| Reasoning | Reasoning | Non-Reasoning |
| Context | 128K | 128K |
| Independent public score (not head-to-head) | 41.19 | null |
| Benchmarks Covered | 8 | 1 |
| Pricing (input/output) | $0.55 / $2.19 | N/A |

## Category Breakdown

### Coding

- Winner: Insufficient shared evidence
- DeepSeek-R1 public-lane score: Coming soon
- DeepSeek-V2.5 public-lane score: Coming soon

| Benchmark | DeepSeek-R1 | DeepSeek-V2.5 | Winner |
|-----------|-----------|-----------|--------|
| SWE-bench Verified | 49.2% | Coming soon | Coming soon |
| HumanEval | Coming soon | 89% | Coming soon |

### Reasoning

- Winner: Insufficient shared evidence
- DeepSeek-R1 public-lane score: 64.1 (Unranked · 2 rankable rows)
- DeepSeek-V2.5 public-lane score: Coming soon

### Knowledge

- Winner: Insufficient shared evidence
- DeepSeek-R1 public-lane score: 41.5 (Supported · #109/181)
- DeepSeek-V2.5 public-lane score: Coming soon

| Benchmark | DeepSeek-R1 | DeepSeek-V2.5 | Winner |
|-----------|-----------|-----------|--------|
| MMLU | 90.8% | Coming soon | Coming soon |
| GPQA | 71.5% | Coming soon | Coming soon |
| MMLU-Pro | 84% | Coming soon | Coming soon |
| SimpleQA | 30.1% | Coming soon | Coming soon |

### Instruction Following

- Winner: Insufficient shared evidence
- DeepSeek-R1 public-lane score: 45.9 (#92/128)
- DeepSeek-V2.5 public-lane score: Coming soon

| Benchmark | DeepSeek-R1 | DeepSeek-V2.5 | Winner |
|-----------|-----------|-----------|--------|
| IFEval | 83.3% | Coming soon | Coming soon |

### Mathematics

- Winner: Insufficient shared evidence
- DeepSeek-R1 public-lane score: Coming soon
- DeepSeek-V2.5 public-lane score: Coming soon

| Benchmark | DeepSeek-R1 | DeepSeek-V2.5 | Winner |
|-----------|-----------|-----------|--------|
| AIME 2024 | 79.8% | Coming soon | Coming soon |
| MATH-500 | 97.3% | Coming soon | Coming soon |

## FAQ

### Can I compare DeepSeek-R1 and DeepSeek-V2.5 on BenchLM yet?

Not fully yet. BenchLM is tracking both models, but sourced benchmark coverage is still incomplete for a fair score-level comparison.

### Why does this page show "coming soon" values?

BenchLM only calls winners when public benchmark coverage is available on both sides of the comparison.

## Related Comparisons

- [DeepSeek-R1 vs DeepSeek-R1-0528](/compare/deepseek-r1-vs-deepseek-r1-0528)
- [DeepSeek-V2.5 vs DeepSeek-R1-0528](/compare/deepseek-r1-0528-vs-deepseek-v2-5)
- [DeepSeek-R1 vs Claude Opus 5.5](/compare/claude-opus-5-5-vs-deepseek-r1)
- [DeepSeek-V2.5 vs Claude Opus 5.5](/compare/claude-opus-5-5-vs-deepseek-v2-5)

## Explore More

- [DeepSeek-R1 profile](/models/deepseek-r1)
- [DeepSeek-V2.5 profile](/models/deepseek-v2-5)
- [Compare Pricing](/llm-pricing)
- [Alternative Finder](/tools/alternative-finder)
- [LLM Selector](/tools/llm-selector)
- [Overall Rankings](/best/overall)