# LFM2.5-VL-3B vs Qwen3.5 397B

> Side-by-side benchmark comparison for LFM2.5-VL-3B and Qwen3.5 397B across agentic, coding, multimodal, knowledge, reasoning, multilingual, and math tasks.

- Canonical page: https://benchlm.ai/compare/lfm2-5-vl-3b-vs-qwen3-5-397b
- Last updated: September 10, 2026

- Shared sourced benchmarks: 2
- HTML indexing: indexable
- Ranking lane: BenchAlign v5

## Quick Verdict

Evidence is too thin for a benchmark verdict: LFM2.5-VL-3B and Qwen3.5 397B 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-3B has the lower published output-token price.
- Qwen3.5 397B has the larger published context window at 128K.

## Model Snapshot

| Property | LFM2.5-VL-3B | Qwen3.5 397B |
|----------|----------|----------|
| Creator | LiquidAI | Alibaba |
| Type | Open Weight | Open Weight |
| Reasoning | Non-Reasoning | Non-Reasoning |
| Context | 32K | 128K |
| Independent public score (not head-to-head) | null | 54.7 |
| Benchmarks Covered | 10 | 38 |
| Pricing (input/output) | $0.00 / $0.00 | $0.60 / $3.60 |

## Category Breakdown

### Agentic

- Winner: Insufficient shared evidence
- LFM2.5-VL-3B public-lane score: Coming soon
- Qwen3.5 397B public-lane score: 46.8 (Estimated · #71/152)

| Benchmark | LFM2.5-VL-3B | Qwen3.5 397B | Winner |
|-----------|-----------|-----------|--------|
| BFCL v4 | 32.5% | Coming soon | Coming soon |
| Terminal-Bench 2.0 | Coming soon | 52.5% | Coming soon |
| BrowseComp | Coming soon | 62% | Coming soon |
| Claw-Eval | Coming soon | 56.8% | Coming soon |
| QwenClawBench | Coming soon | 51.8% | Coming soon |
| τ³-bench results | Coming soon | 68.4% | Coming soon |
| VITA-Bench | Coming soon | 43.7% | Coming soon |
| DeepPlanning | Coming soon | 37.6% | Coming soon |
| Toolathlon | Coming soon | 36.3% | Coming soon |
| MCP Atlas | Coming soon | 46.1% | Coming soon |
| MCP-Tasks | Coming soon | 74.2% | Coming soon |
| WideResearch | Coming soon | 74.0% | Coming soon |
| Gert Labs | Coming soon | 46.76% | Coming soon |
| ResearchClawBench | Coming soon | 14.2% | Coming soon |

### Coding

- Winner: Insufficient shared evidence
- LFM2.5-VL-3B public-lane score: Coming soon
- Qwen3.5 397B public-lane score: 50.2 (Estimated · #59/151)

| Benchmark | LFM2.5-VL-3B | Qwen3.5 397B | Winner |
|-----------|-----------|-----------|--------|
| SWE-bench Verified | Coming soon | 76.2% | Coming soon |
| LiveCodeBench v6 | Coming soon | 83.6% | Coming soon |
| SWE-bench Pro | Coming soon | 50.9% | Coming soon |

### Multimodal & Grounded

- Winner: Insufficient shared evidence
- LFM2.5-VL-3B public-lane score: Coming soon
- Qwen3.5 397B public-lane score: 63 (#27/48)

| Benchmark | LFM2.5-VL-3B | Qwen3.5 397B | Winner |
|-----------|-----------|-----------|--------|
| RealWorldQA | 73.1% | Coming soon | Coming soon |
| SimpleVQA | 35.4% | Coming soon | Coming soon |
| CountBench | 87.3% | Coming soon | Coming soon |
| MMMU | 48.4% | Coming soon | Coming soon |
| MMMU-Pro | 30.5% | 79% | Qwen3.5 397B |
| OCRBench V2 | 47.5% | Coming soon | Coming soon |
| RefCOCO (avg) | 87.9% | Coming soon | Coming soon |
| MathVision | Coming soon | 88.6% | Coming soon |
| CharXiv | Coming soon | 80.8% | Coming soon |
| VideoMMMU | Coming soon | 84.7% | Coming soon |
| ScreenSpot Pro | Coming soon | 65.6% | Coming soon |
| V* | Coming soon | 95.8% | Coming soon |

### Reasoning

- Winner: Insufficient shared evidence
- LFM2.5-VL-3B public-lane score: Coming soon
- Qwen3.5 397B public-lane score: 59.8 (Unranked · 2 rankable rows)

| Benchmark | LFM2.5-VL-3B | Qwen3.5 397B | Winner |
|-----------|-----------|-----------|--------|
| LongBench v2 | Coming soon | 63.2% | Coming soon |
| AI-Needle | Coming soon | 68.7% | Coming soon |

### Knowledge

- Winner: Insufficient shared evidence
- LFM2.5-VL-3B public-lane score: Coming soon
- Qwen3.5 397B public-lane score: 50.2 (Estimated · #77/183)

| Benchmark | LFM2.5-VL-3B | Qwen3.5 397B | Winner |
|-----------|-----------|-----------|--------|
| GPQA | Coming soon | 88.4% | Coming soon |
| SuperGPQA | Coming soon | 70.4% | Coming soon |
| MMLU-Pro | Coming soon | 87.8% | Coming soon |
| MMLU-Redux | Coming soon | 94.9% | Coming soon |
| C-Eval | Coming soon | 93% | Coming soon |
| HLE | Coming soon | 28.7% | Coming soon |

### Instruction Following

- Winner: Insufficient shared evidence
- LFM2.5-VL-3B public-lane score: Coming soon
- Qwen3.5 397B public-lane score: 0 (#123/123)

| Benchmark | LFM2.5-VL-3B | Qwen3.5 397B | Winner |
|-----------|-----------|-----------|--------|
| IFEval | 82.3% | 92.6% | Qwen3.5 397B |
| IFBench | 25.8% | Coming soon | Coming soon |

### Multilingual

- Winner: Insufficient shared evidence
- LFM2.5-VL-3B public-lane score: Coming soon
- Qwen3.5 397B public-lane score: 69.7 (#5/12)

| Benchmark | LFM2.5-VL-3B | Qwen3.5 397B | Winner |
|-----------|-----------|-----------|--------|
| MMLU-ProX | Coming soon | 84.7% | Coming soon |
| NOVA-63 | Coming soon | 59.1% | Coming soon |

### Mathematics

- Winner: Insufficient shared evidence
- LFM2.5-VL-3B public-lane score: Coming soon
- Qwen3.5 397B public-lane score: 74.2 (Unranked · 5 rankable rows)

| Benchmark | LFM2.5-VL-3B | Qwen3.5 397B | Winner |
|-----------|-----------|-----------|--------|
| AIME26 | Coming soon | 93.3% | Coming soon |
| HMMT Feb 2025 | Coming soon | 94.8% | Coming soon |
| HMMT Nov 2025 | Coming soon | 92.7% | Coming soon |
| HMMT Feb 2026 | Coming soon | 87.9% | Coming soon |
| MMAnswerBench | Coming soon | 80.9% | Coming soon |

## FAQ

### Is there enough evidence to pick between LFM2.5-VL-3B and Qwen3.5 397B?

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

- [LFM2.5-VL-3B vs Qwen3.5 397B (Reasoning)](/compare/lfm2-5-vl-3b-vs-qwen3-5-397b-reasoning)
- [Qwen3.5 397B vs Qwen3.5 397B (Reasoning)](/compare/qwen3-5-397b-vs-qwen3-5-397b-reasoning)
- [LFM2.5-VL-3B vs Claude Fable 5.1](/compare/claude-fable-5-1-vs-lfm2-5-vl-3b)
- [Qwen3.5 397B vs Claude Fable 5.1](/compare/claude-fable-5-1-vs-qwen3-5-397b)

## Explore More

- [LFM2.5-VL-3B profile](/models/lfm2-5-vl-3b)
- [Qwen3.5 397B profile](/models/qwen3-5-397b)
- [Compare Pricing](/llm-pricing)
- [Alternative Finder](/tools/alternative-finder)
- [LLM Selector](/tools/llm-selector)
- [Overall Rankings](/best/overall)