Pricing
Input -
Output -
Text to Image -
Image to Image -
Text to Video -
Image to Video -
Capabilities
Input Types Text / Image / Video
Output Types Text
Function Calling Supported
Session Cache -
Fine-Tuning -
Prefix Cache -
Structured Output Supported
Image Understanding Supported
Text Generation Supported
Text to Image -
Image to Image -
Text to Video -
Image to Video -
Open Source Status Open source
Parameters
Last Updated 2026-08-08
Context Length -
Max Input Tokens 32K
Max Output Tokens 32K
Max Thought Tokens -
TPM -
RPM -
Parameter Count 1.9B params
Modalities Multimodal
Language Support Multilingual
Best For General AI application workflows
Resolution -
Video Duration -
Frame Rate -
Benchmarks
These benchmark scores summarize how the model performs across different tasks and evaluation sets.
AIME
11.6 Measures performance on difficult competition-level mathematics problems.
GPQA
23.7 Evaluates graduate-level science knowledge and reasoning ability.
LiveCodeBench
13.2 Benchmarks code understanding and real-world coding task performance.
HumanEval
75 Evaluates code generation correctness on programming tasks.
MMLU Pro
50.6 Tests multidisciplinary knowledge, comprehension, and reasoning.