Liquid AI model product
LFM2.5 VL 3B
12B parameters; HF safetensors total 3,123,483,888) for on-device and edge image-text→text.
Updated Aug 17, 2026. Default version: LFM2.5-VL-3B
LLMBoard Score6.5LFM2.5-VL-3B
Coverage40%22 benchmark families
Context windowN/ATokens
Official input priceN/AOfficial price unavailable
LFM2.5 VL 3B Capability Profile
This profile uses the model's current scored version. Arena ratings and prices are shown separately.
LFM2.5 VL 3B Benchmark Results
Benchmark scores for LFM2.5-VL-3B.
Sort by
Benchmark | Score | Rank | Participants | Percentile | Evidence | Evaluated |
|---|
| BenchmarkMME | Score73.1% | Rank01 | Participants4 | Percentile100.0% | EvidenceC | Evaluated |
| BenchmarkMM IF-Eval | Score60.6% | Rank01 | Participants2 | Percentile100.0% | EvidenceC | Evaluated |
| BenchmarkPOPE | Score88.7% | Rank01 | Participants3 | Percentile100.0% | EvidenceC | Evaluated |
| BenchmarkTextVQA | Score84.3% | Rank04 | Participants16 | Percentile80.0% | EvidenceC | Evaluated |
| BenchmarkRefCOCO-avg | Score0.879 points | Rank08 | Participants9 | Percentile12.5% | EvidenceC | Evaluated |
| BenchmarkMMMU (val) | Score48.4% | Rank12 | Participants13 | Percentile8.3% | EvidenceC | Evaluated |
| BenchmarkMuirBench | Score58.3% | Rank12 | Participants12 | Percentile0.0% | EvidenceC | Evaluated |
| BenchmarkBLINK | Score61.5% | Rank13 | Participants15 | Percentile14.3% | EvidenceC | Evaluated |
| BenchmarkOCRBench-V2 (en) | Score47.5% | Rank13 | Participants14 | Percentile7.7% | EvidenceC | Evaluated |
| BenchmarkBFCL-V4 | Score32.5% | Rank14 | Participants15 | Percentile7.1% | EvidenceC | Evaluated |
| BenchmarkSimpleVQA | Score0.354 points | Rank14 | Participants14 | Percentile0.0% | EvidenceC | Evaluated |
| BenchmarkMMBench-V1.1 | Score81.0% | Rank16 | Participants20 | Percentile21.1% | EvidenceC | Evaluated |
| BenchmarkOCRBench | Score84.2% | Rank17 | Participants24 | Percentile30.4% | EvidenceC | Evaluated |
| BenchmarkHallusion Bench | Score47.2% | Rank18 | Participants18 | Percentile0.0% | EvidenceC | Evaluated |
| BenchmarkDocVQA | Score91.1% | Rank19 | Participants28 | Percentile33.3% | EvidenceC | Evaluated |
| BenchmarkChartQA | Score81.3% | Rank20 | Participants26 | Percentile24.0% | EvidenceC | Evaluated |
| BenchmarkMathVista-Mini | Score68.5% | Rank20 | Participants24 | Percentile17.4% | EvidenceC | Evaluated |
| BenchmarkMMStar | Score63.3% | Rank20 | Participants24 | Percentile17.4% | EvidenceC | Evaluated |
| BenchmarkMulti-IF | Score59.4% | Rank21 | Participants23 | Percentile9.1% | EvidenceC | Evaluated |
| BenchmarkRealWorldQA | Score73.1% | Rank21 | Participants29 | Percentile28.6% | EvidenceC | Evaluated |
| BenchmarkIFBench | Score25.8% | Rank33 | Participants34 | Percentile3.0% | EvidenceC | Evaluated |
| BenchmarkIFEval | Score82.3% | Rank49 | Participants67 | Percentile27.3% | EvidenceC | Evaluated |
| BenchmarkMMMU-Pro | Score30.5% | Rank68 | Participants68 | Percentile0.0% | EvidenceC | Evaluated |
LFM2.5 VL 3B Arena Results
Preference and agent-evaluation results for the default version.
No Arena results
The default version does not have a matching Arena result yet.
LFM2.5 VL 3B Pricing
Official vendor API pricing appears first, followed by individual provider offers.
- Official API
- N/A
- Official provider
- N/A
- Lowest third-party
- N/A
- Tracked offerings
- 0
No provider prices
The default version has no current input or output token prices.
Official prices use only the vendor's configured official Provider and positive standard USD PAYG rates. Third-party offers remain explicitly labeled.
LFM2.5 VL 3B Runtime Performance
Provider-specific output speed and catalog latency for LFM2.5-VL-3B. Runtime does not affect the capability score.
No runtime data
Browse runtime rankingsNo provider-specific speed or latency record is linked to the default version yet.
Output Speed is generated output tokens received per second. Catalog latency is reported separately from observed provider TTFT.
LFM2.5 VL 3B Specifications
Technical details for the model's default version.
- Version
- LFM2.5-VL-3B
- Released
- Aug 12, 2026
- Knowledge cutoff
- Unknown
- Parameters
- 3.1B
- Context window
- N/A
- Max output
- N/A
- Inputs
- image, text
- Outputs
- text
- Open weights
- No
- License
- LFM Open License v1.0
LFM2.5 VL 3B Versions
Available versions of this model. The score column identifies the version used in the overall ranking.
Sort by
Version | Released | LLMBoard | Parameters | Context | Max output | Open weights | License |
|---|
| VersionLFM2.5-VL-3B | Released | LLMBoard6.5 | Parameters3.1B | ContextN/A | Max outputN/A | Open weightsNo | LicenseLFM Open License v1.0 |
LFM2.5 VL 3B vs nearby models
Open a comparison with the three ranked models immediately above and below this model.
Models similar to LFM2.5 VL 3B
Recommendations prioritize the same model type and family, then the closest LLMBoard score.
LA
LFM2.5 2.6B
Liquid AI
MA
Mistral Large 2
Mistral AI
MA
Mistral Small 3 24B
Mistral AI
NR
Hermes 3 70B
Nous Research
MI
Phi 4 multimodal
Microsoft
AC
Qwen2.5 7B
Alibaba Cloud / Qwen Team
What is LFM2.5 VL 3B?
Key information about LFM2.5 VL 3B and its available data.
12B parameters; HF safetensors total 3,123,483,888) for on-device and edge image-text→text. Non-reasoning: answers directly for low latency. 6B language backbone with a SigLIP2 400M NaFlex vision encoder; improves screen/UI understanding, grounding, function calling, and multi-image input over LFM2-VL-3B.
Pre-trained on ~34T tokens; 128K vocabulary; 32,768-token context. cpp, MLX, vLLM, SGLang, and ONNX. 0). 1, MM-IF-Eval, MathVista-Mini, MMMU-Pro, MMMU (val), ChartQA, DocVQA, OCRBench, OCRBench-V2 (en), TextVQA, RefCOCO-avg, BLINK, MuirBench, HallusionBench, POPE, IFEval, IFBench, Multi-IF, and BFCL-v4.
7 avg), CountBenchQA, SEED-Bench (image), MMMB, Multilingual MMBench, LogicVista, InfographicVQA, or ToolSandbox — recorded here only. 5-VL-3B. 6b:free for the text sibling).
Data as of 2026-08-17.
FAQ
Common questions about LFM2.5 VL 3B.
When was LFM2.5 VL 3B released?
LFM2.5 VL 3B's default version was released on Aug 12, 2026.
How much does LFM2.5 VL 3B cost?
No official standard PAYG price is currently available for LFM2.5 VL 3B.
Who created LFM2.5 VL 3B?
LFM2.5 VL 3B was created by Liquid AI.
What is the context window for LFM2.5 VL 3B?
A context window is not available for the default version.
Is LFM2.5 VL 3B open weight?
No. The default version is not marked as having publicly available weights.
How many API providers offer LFM2.5 VL 3B?
No provider offering is currently linked to the default version.
What models should I compare LFM2.5 VL 3B with?
Nearby ranked alternatives include GPT-4o-mini, Hermes 3 70B, Mistral Small 3 24B.