# LFM3-VL-2B vs Step 5 Preview

> Side-by-side benchmark comparison for LFM3-VL-2B and Step 5 Preview across agentic, coding, multimodal, knowledge, reasoning, multilingual, and math tasks.

- Canonical page: https://benchlm.ai/compare/lfm3-vl-2b-vs-step-5-preview
- Last updated: October 27, 2026

- Shared sourced benchmarks: 1
- HTML indexing: indexable
- Ranking lane: BenchAlign v5.8

## Quick Verdict

Evidence is too thin for a benchmark verdict: LFM3-VL-2B and Step 5 Preview share only 1 sourced benchmark row.

## Summary

- Only 1 shared sourced benchmark row is available. That is not enough for a broad capability verdict, so this mirror does not name an overall winner.
- LFM3-VL-2B has the lower published output-token price.
- Step 5 Preview has the larger published context window at 1M.

## Model Snapshot

| Property | LFM3-VL-2B | Step 5 Preview |
|----------|----------|----------|
| Creator | LiquidAI | StepFun |
| Type | Open Weight | Open Weight |
| Reasoning | Reasoning | Reasoning |
| Context | 128K | 1M |
| Independent public score (not head-to-head) | null | 69.51 |
| Benchmarks Covered | 20 | 24 |
| Pricing (input/output) | $0.00 / $0.00 | $1.00 / $2.70 |

## Category Breakdown

### Agentic

- Winner: Insufficient shared evidence
- LFM3-VL-2B public-lane score: Coming soon
- Step 5 Preview public-lane score: 65.6 (Estimated · #18/134)

| Benchmark | LFM3-VL-2B | Step 5 Preview | Winner |
|-----------|-----------|-----------|--------|
| Terminal-Bench 2.1 | Coming soon | 85.0% | Coming soon |
| Terminal-Bench 4.0 | Coming soon | 33.30% | Coming soon |
| CyberGym | Coming soon | 84.7% | Coming soon |
| AutomationBench | Coming soon | 44.0% | Coming soon |
| Toolathlon-Verified | Coming soon | 74.1% | Coming soon |
| MCP Atlas | Coming soon | 85.6% | Coming soon |
| JobBench | Coming soon | 59.0% | Coming soon |
| APEX-Agents | Coming soon | 37.8% | Coming soon |
| DRACO | Coming soon | 83.3% | Coming soon |
| BrowseComp | Coming soon | 88.7% | Coming soon |
| Agents' Last Exam | Coming soon | 29.5% | Coming soon |
| Hermes Index | Coming soon | 32.8% | Coming soon |
| Hermes Bench | Coming soon | 69.2% | Coming soon |

### Coding

- Winner: Insufficient shared evidence
- LFM3-VL-2B public-lane score: Coming soon
- Step 5 Preview public-lane score: 60.7 (Estimated · #20/155)

| Benchmark | LFM3-VL-2B | Step 5 Preview | Winner |
|-----------|-----------|-----------|--------|
| DeepSWE | Coming soon | 67.7% | Coming soon |
| Terminal-Bench 2.1 | Coming soon | 85.0% | Coming soon |
| SciCode | Coming soon | 58.9% | Coming soon |
| ProgramBench | Coming soon | 80.5% | Coming soon |
| sweMarathon | Coming soon | 72.7% | Coming soon |
| MLS-Bench Lite | Coming soon | 40.5% | Coming soon |

### Reasoning

- Winner: Insufficient shared evidence
- LFM3-VL-2B public-lane score: Coming soon
- Step 5 Preview public-lane score: 82.5 (Unranked · 2 rankable rows)

| Benchmark | LFM3-VL-2B | Step 5 Preview | Winner |
|-----------|-----------|-----------|--------|
| CritPt | Coming soon | 20.9% | Coming soon |

### Multimodal & Grounded

- Winner: Insufficient shared evidence
- LFM3-VL-2B public-lane score: 19.6 (Unranked · 19 rankable rows)
- Step 5 Preview public-lane score: 73.4 (#32/58)

| Benchmark | LFM3-VL-2B | Step 5 Preview | Winner |
|-----------|-----------|-----------|--------|
| RealWorldQA | 79.3% | Coming soon | Coming soon |
| SimpleVQA | 36.3% | Coming soon | Coming soon |
| CountBench | 93.0% | Coming soon | Coming soon |
| MMMU | 55.7% | Coming soon | Coming soon |
| MMMU-Pro | 45.7% | 76% | Step 5 Preview |
| OCRBench V2 | 63.5% | Coming soon | Coming soon |
| RefCOCO (avg) | 90.5% | Coming soon | Coming soon |
| V* | 86.4% | Coming soon | Coming soon |
| CharXiv w/o tools | 70.2% | Coming soon | Coming soon |
| MathVision | 51.6% | Coming soon | Coming soon |
| DynaMath | 79.2% | Coming soon | Coming soon |
| We-Math | 68.6% | Coming soon | Coming soon |
| ERQA | 48.2% | Coming soon | Coming soon |
| ScreenSpot Pro | 55.5% | Coming soon | Coming soon |
| LVBench | 41.2% | Coming soon | Coming soon |
| MLVU (M-Avg) | 71.3% | Coming soon | Coming soon |
| MMVU | 52.9% | Coming soon | Coming soon |
| Video-MME (w/o subtitle) | 62.8% | Coming soon | Coming soon |
| Video-MME (with subtitle) | 69.0% | Coming soon | Coming soon |
| VideoMMMU | 49.6% | Coming soon | Coming soon |
| OfficeQA Pro | Coming soon | 60.3% | Coming soon |

### Knowledge

- Winner: Insufficient shared evidence
- LFM3-VL-2B public-lane score: Coming soon
- Step 5 Preview public-lane score: 68.4 (Supported · #23/189)

| Benchmark | LFM3-VL-2B | Step 5 Preview | Winner |
|-----------|-----------|-----------|--------|
| GPQA-D | Coming soon | 93.5% | Coming soon |
| HLE | Coming soon | 46.5% | Coming soon |

## FAQ

### Is there enough evidence to pick between LFM3-VL-2B and Step 5 Preview?

Not yet. The models share only 1 sourced benchmark row, 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

- [LFM3-VL-2B vs Claude Opus 5.5](/compare/claude-opus-5-5-vs-lfm3-vl-2b)
- [Step 5 Preview vs Claude Opus 5.5](/compare/claude-opus-5-5-vs-step-5-preview)
- [LFM3-VL-2B vs GPT-6 Astra](/compare/gpt-6-astra-vs-lfm3-vl-2b)
- [Step 5 Preview vs GPT-6 Astra](/compare/gpt-6-astra-vs-step-5-preview)

## Explore More

- [LFM3-VL-2B profile](/models/lfm3-vl-2b)
- [Step 5 Preview profile](/models/step-5-preview)
- [Compare Pricing](/llm-pricing)
- [Alternative Finder](/tools/alternative-finder)
- [LLM Selector](/tools/llm-selector)
- [Overall Rankings](/best/overall)