DE
DeepSeek-R1-Distill-Llama
DeepSeek
DeepSeek model product
DeepSeek-VL2
An advanced series of large Mixture-of-Experts (MoE) Vision-Language Models that significantly improves upon its predecessor, DeepSeek-VL.
Updated Aug 17, 2026. Default version: DeepSeek VL2 Tiny
LLMBoard Score0.0DeepSeek VL2 Tiny
Coverage0%14 benchmark families
Context windowN/ATokens
Official input priceN/AOfficial price unavailable
DeepSeek-VL2 Benchmark Results
Benchmark scores for DeepSeek VL2 Tiny.
Sort by
Benchmark | Score | Rank | Participants | Percentile | Evidence | Evaluated |
|---|
| BenchmarkMME | Score19.1% | Rank04 | Participants4 | Percentile0.0% | EvidenceC | Evaluated |
| BenchmarkMMT-Bench | Score53.2% | Rank04 | Participants4 | Percentile0.0% | EvidenceC | Evaluated |
| BenchmarkInfoVQA | Score66.1% | Rank07 | Participants10 | Percentile33.3% | EvidenceC | Evaluated |
| BenchmarkTextVQA | Score80.7% | Rank08 | Participants16 | Percentile53.3% | EvidenceC | Evaluated |
| BenchmarkMMBench | Score69.2% | Rank09 | Participants9 | Percentile0.0% | EvidenceC | Evaluated |
| BenchmarkMMBench-V1.1 | Score68.3% | Rank20 | Participants20 | Percentile0.0% | EvidenceC | Evaluated |
| BenchmarkChartQA | Score81.0% | Rank21 | Participants26 | Percentile20.0% | EvidenceC | Evaluated |
| BenchmarkDocVQA | Score88.9% | Rank22 | Participants28 | Percentile22.2% | EvidenceC | Evaluated |
| BenchmarkOCRBench | Score80.9% | Rank22 | Participants24 | Percentile8.7% | EvidenceC | Evaluated |
| BenchmarkMMStar | Score45.9% | Rank24 | Participants24 | Percentile0.0% | EvidenceC | Evaluated |
| BenchmarkRealWorldQA | Score64.2% | Rank28 | Participants29 | Percentile3.6% | EvidenceC | Evaluated |
| BenchmarkAI2D | Score71.6% | Rank33 | Participants33 | Percentile0.0% | EvidenceC | Evaluated |
| BenchmarkMathVista | Score53.6% | Rank33 | Participants39 | Percentile15.8% | EvidenceC | Evaluated |
| BenchmarkMMMU | Score40.7% | Rank62 | Participants63 | Percentile1.6% | EvidenceC | Evaluated |
DeepSeek-VL2 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.
DeepSeek-VL2 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.
DeepSeek-VL2 Runtime Performance
Provider-specific output speed and catalog latency for DeepSeek VL2 Tiny. 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.
DeepSeek-VL2 Specifications
Technical details for the model's default version.
- Version
- DeepSeek VL2 Tiny
- Released
- Dec 13, 2024
- Knowledge cutoff
- Unknown
- Parameters
- 3B
- Context window
- N/A
- Max output
- N/A
- Inputs
- image, text
- Outputs
- text
- Open weights
- No
- License
- deepseek
DeepSeek-VL2 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 |
|---|
| VersionDeepSeek VL2 | Released | LLMBoardN/A | Parameters27B | Context129.3K | Max output129.3K | Open weightsNo | Licensedeepseek |
| VersionDeepSeek VL2 Small | Released | LLMBoardN/A | Parameters16B | ContextN/A | Max outputN/A | Open weightsNo | Licensedeepseek |
| VersionDeepSeek VL2 Tiny | Released | LLMBoard0.0 | Parameters3B | ContextN/A | Max outputN/A | Open weightsNo | Licensedeepseek |
DeepSeek-VL2 vs nearby models
Open a comparison with the three ranked models immediately above and below this model.
Models similar to DeepSeek-VL2
Recommendations prioritize the same model type and family, then the closest LLMBoard score.
DE
DeepSeek-V2.5
DeepSeek
DE
DeepSeek-R1-Distill-Qwen
DeepSeek
DE
DeepSeek-V3
DeepSeek
DE
DeepSeek-V3.1
DeepSeek
DE
DeepSeek-R1
DeepSeek
What is DeepSeek-VL2?
Key information about DeepSeek-VL2 and its available data.
An advanced series of large Mixture-of-Experts (MoE) Vision-Language Models that significantly improves upon its predecessor, DeepSeek-VL.
DeepSeek-VL2 demonstrates superior capabilities across various tasks, including but not limited to visual question answering, optical character recognition, document/table/chart understanding, and visual grounding.
Data as of 2026-08-17.
FAQ
Common questions about DeepSeek-VL2.
When was DeepSeek-VL2 released?
DeepSeek-VL2's default version was released on Dec 13, 2024.
How much does DeepSeek-VL2 cost?
No official standard PAYG price is currently available for DeepSeek-VL2.
Who created DeepSeek-VL2?
DeepSeek-VL2 was created by DeepSeek.
What is the context window for DeepSeek-VL2?
A context window is not available for the default version.
Is DeepSeek-VL2 open weight?
No. The default version is not marked as having publicly available weights.
How many API providers offer DeepSeek-VL2?
No provider offering is currently linked to the default version.
What models should I compare DeepSeek-VL2 with?
Nearby ranked alternatives include Llama 3.1 Nemotron 70B, Jamba 1.5 Large, Codestral 22B.