# DeepSeek V4 Pro vs LFM2.5-VL-450M

> Side-by-side benchmark comparison for DeepSeek V4 Pro and LFM2.5-VL-450M across agentic, coding, multimodal, knowledge, reasoning, multilingual, and math tasks.

- Canonical page: https://benchlm.ai/compare/deepseek-v4-pro-vs-lfm2-5-vl-450m
- Last updated: November 19, 2026

- Shared sourced benchmarks: 2
- HTML indexing: indexable
- Ranking lane: BenchAlign v6.0

## Quick Verdict

Evidence is too thin for a benchmark verdict: DeepSeek V4 Pro and LFM2.5-VL-450M share only 2 sourced benchmark rows.

## Summary

- Only 2 shared sourced benchmark rows are available. That is not enough for a broad capability verdict, so this mirror does not name an overall winner.
- LFM2.5-VL-450M has the lower published output-token price.
- DeepSeek V4 Pro has the larger published context window at 1M.

## Model Snapshot

| Property | DeepSeek V4 Pro | LFM2.5-VL-450M |
|----------|----------|----------|
| Creator | DeepSeek | LiquidAI |
| Type | Open Weight | Open Weight |
| Reasoning | Reasoning | Non-Reasoning |
| Context | 1M | 128K |
| Independent public score (not head-to-head) | 61.5 | null |
| Benchmarks Covered | 45 | 7 |
| Pricing (input/output) | $1.32 / $3.96 | $0.00 / $0.00 |

## Category Breakdown

### Agentic

- Winner: Insufficient shared evidence
- DeepSeek V4 Pro public-lane score: 59.3 (Supported · #60/177)
- LFM2.5-VL-450M public-lane score: Coming soon

| Benchmark | DeepSeek V4 Pro | LFM2.5-VL-450M | Winner |
|-----------|-----------|-----------|--------|
| Terminal-Bench 4.1 | 6.70% | Coming soon | Coming soon |
| Finch | 67.4% | Coming soon | Coming soon |
| Terminal-Bench 2.0 | 67.9% | Coming soon | Coming soon |
| Terminal-Bench 2.1 | 87.9% | Coming soon | Coming soon |
| BrowseComp | 83.4% | Coming soon | Coming soon |
| HLE w/ tools | 60.0% | Coming soon | Coming soon |
| MCP Atlas | 73.6% | Coming soon | Coming soon |
| Toolathlon | 51.8% | Coming soon | Coming soon |
| CyberGym | 83.3% | Coming soon | Coming soon |
| Toolathlon-Verified | 74.1% | Coming soon | Coming soon |
| Agents' Last Exam | 25.7% | Coming soon | Coming soon |
| AutomationBench | 31.8% | Coming soon | Coming soon |
| Terminal-Bench 2.1 (Vals) | 54.7% | Coming soon | Coming soon |
| BFCL v4 | Coming soon | 21.1% | Coming soon |

### Coding

- Winner: Insufficient shared evidence
- DeepSeek V4 Pro public-lane score: 59.6 (Supported · #55/177)
- LFM2.5-VL-450M public-lane score: Coming soon

| Benchmark | DeepSeek V4 Pro | LFM2.5-VL-450M | Winner |
|-----------|-----------|-----------|--------|
| LiveCodeBench Pass@1-COT | 93.5% | Coming soon | Coming soon |
| Codeforces | 3206.0 | Coming soon | Coming soon |
| SWE-bench Verified | 80.6% | Coming soon | Coming soon |
| SWE-bench Pro | 55.4% | Coming soon | Coming soon |
| SWE Multilingual | 76.2% | Coming soon | Coming soon |
| Terminal-Bench 2.0 | 67.9% | Coming soon | Coming soon |
| Vibe Code Bench | 49.93% | Coming soon | Coming soon |
| Terminal-Bench 2.1 | 87.9% | Coming soon | Coming soon |
| NL2Repo | 61.5% | Coming soon | Coming soon |
| DeepSWE | 62.7% | Coming soon | Coming soon |
| DSBench-FullStack | 71.1% | Coming soon | Coming soon |
| DSBench-Hard | 67.2% | Coming soon | Coming soon |
| OpenHarmony Bench v1.0 | 59.0% | Coming soon | Coming soon |
| LiveCodeBench (Vals) | 87.5% | Coming soon | Coming soon |
| SWE-bench (Vals) | 96.4% | Coming soon | Coming soon |

### Reasoning

- Winner: Insufficient shared evidence
- DeepSeek V4 Pro public-lane score: 64.6 (Supported · #33/160)
- LFM2.5-VL-450M public-lane score: Coming soon

| Benchmark | DeepSeek V4 Pro | LFM2.5-VL-450M | Winner |
|-----------|-----------|-----------|--------|
| MRCR 1M | 83.5% | Coming soon | Coming soon |
| CorpusQA 1M | 62.0% | Coming soon | Coming soon |
| ARC-AGI-1 | 90.00% | Coming soon | Coming soon |
| ARC-AGI-2 | 61.3% | Coming soon | Coming soon |

### Multimodal & Grounded

- Winner: Insufficient shared evidence
- DeepSeek V4 Pro public-lane score: 60.6 (Estimated · #51/115)
- LFM2.5-VL-450M public-lane score: Coming soon

| Benchmark | DeepSeek V4 Pro | LFM2.5-VL-450M | Winner |
|-----------|-----------|-----------|--------|
| MMMU | Coming soon | 32.7% | Coming soon |
| RealWorldQA | Coming soon | 58.4% | Coming soon |
| CountBench | Coming soon | 73.3% | Coming soon |

### Knowledge

- Winner: Insufficient shared evidence
- DeepSeek V4 Pro public-lane score: 60.2 (Supported · #44/156)
- LFM2.5-VL-450M public-lane score: Coming soon

| Benchmark | DeepSeek V4 Pro | LFM2.5-VL-450M | Winner |
|-----------|-----------|-----------|--------|
| MMLU-Pro | 87.5% | 19.3% | DeepSeek V4 Pro |
| Chinese-SimpleQA | 84.4% | Coming soon | Coming soon |
| GPQA | 90.1% | 25.7% | DeepSeek V4 Pro |
| GPQA-D | 90.1% | Coming soon | Coming soon |
| HLE | 42.7% | Coming soon | Coming soon |
| GPQA Diamond (Vals) | 92.4% | Coming soon | Coming soon |
| MMLU-Pro (Vals) | 87.0% | Coming soon | Coming soon |

### Multilingual

- Winner: Insufficient shared evidence
- DeepSeek V4 Pro public-lane score: 57.5 (Estimated · #31/86)
- LFM2.5-VL-450M public-lane score: Coming soon

### Instruction Following

- Winner: Insufficient shared evidence
- DeepSeek V4 Pro public-lane score: 56.5 (Supported · #46/142)
- LFM2.5-VL-450M public-lane score: Coming soon

| Benchmark | DeepSeek V4 Pro | LFM2.5-VL-450M | Winner |
|-----------|-----------|-----------|--------|
| IFEval | Coming soon | 61.2% | Coming soon |

### Mathematics

- Winner: Insufficient shared evidence
- DeepSeek V4 Pro public-lane score: 59.5 (Supported · #33/124)
- LFM2.5-VL-450M public-lane score: Coming soon

| Benchmark | DeepSeek V4 Pro | LFM2.5-VL-450M | Winner |
|-----------|-----------|-----------|--------|
| HMMT Feb 2026 | 95.2% | Coming soon | Coming soon |
| IMOAnswerBench | 89.8% | Coming soon | Coming soon |
| Apex | 38.3% | Coming soon | Coming soon |
| Apex Shortlist | 90.2% | Coming soon | Coming soon |
| FrontierMath v2 (Tiers 1-3) | 64.561% | Coming soon | Coming soon |
| FrontierMath v2 (Tier 4) | 26.829% | Coming soon | Coming soon |

## FAQ

### Is there enough evidence to pick between DeepSeek V4 Pro and LFM2.5-VL-450M?

Not yet. The models share only 2 sourced benchmark rows, so this mirror reports the available measurements and metadata without naming a broad winner.

### What can I compare safely on this page?

Published pricing, context windows, model type, and individual shared benchmark rows are directly comparable. Independent overall scores and non-overlapping benchmark coverage are labeled separately and should not be read as a controlled head-to-head result.

## Related Comparisons

- [DeepSeek V4 Pro vs Claude Fable 5.2](/compare/claude-fable-5-2-vs-deepseek-v4-pro)
- [LFM2.5-VL-450M vs Claude Fable 5.2](/compare/claude-fable-5-2-vs-lfm2-5-vl-450m)
- [DeepSeek V4 Pro vs GPT-6.2 Astra](/compare/deepseek-v4-pro-vs-gpt-6-2-astra)
- [LFM2.5-VL-450M vs GPT-6.2 Astra](/compare/gpt-6-2-astra-vs-lfm2-5-vl-450m)

## Explore More

- [DeepSeek V4 Pro profile](/models/deepseek-v4-pro)
- [LFM2.5-VL-450M profile](/models/lfm2-5-vl-450m)
- [Compare Pricing](/llm-pricing)
- [Alternative Finder](/tools/alternative-finder)
- [LLM Selector](/tools/llm-selector)
- [Overall Rankings](/best/overall)
