Google DeepMind releases Nano Banana 2 Lite at $0.034 per 1K image with 4-second generation, opens Gemini Omni Flash API
Google DeepMind released Nano Banana 2 Lite (gemini-3.1-flash-lite-image), its fastest image generation model with 4-second text-to-image latency priced at $0.034 per 1K-resolution image. The company also opened developer access to Gemini Omni Flash (gemini-omni-flash-preview) for video generation and editing at $0.10 per second of output.
Google DeepMind releases Nano Banana 2 Lite at $0.034 per 1K image with 4-second generation, opens Gemini Omni Flash API at $0.10 per video second
Google DeepMind released Nano Banana 2 Lite (gemini-3.1-flash-lite-image), its fastest image generation model with 4-second text-to-image latency priced at $0.034 per 1K-resolution image. The company also opened developer access to Gemini Omni Flash (gemini-omni-flash-preview) for video generation and editing at $0.10 per second of output.
Nano Banana 2 Lite targets high-throughput developer workflows where speed and cost matter more than maximum quality. According to Google, it's positioned as the recommended replacement for the original Nano Banana (gemini-2.5-flash-image).
Nano Banana 2 Lite specifications
- Latency: 4 seconds for text-to-image generation
- Pricing: $0.034 per 1K-resolution image
- API name: gemini-3.1-flash-lite-image
- Availability: Google AI Studio, Gemini API, Gemini Enterprise Agent Platform
Google claims the model maintains "reliable prompt adherence, strong character consistency and legible in-image text rendering" despite optimizations for speed. The company published internal benchmarks comparing Elo quality scores, latency, and cost against competitor models, though independent verification is not yet available.
The Nano Banana family now includes four models:
- Nano Banana 2 Lite (Gemini 3.1 Flash Lite Image): Speed-optimized
- Nano Banana 2 (Gemini 3.1 Flash Image): Balanced performance
- Nano Banana Pro (Gemini 3 Pro Image): Quality-optimized for professional use
- Nano Banana (Gemini 2.5 Flash Image): Legacy model, upgrade recommended
Gemini Omni Flash video capabilities
Gemini Omni Flash, first introduced at Google I/O, is now accessible via API in public preview. The model handles video generation and editing from text, image, and video inputs with conversational refinement.
Key specifications:
- Pricing: $0.10 per second of video output (matching Veo 3.1 Fast)
- API name: gemini-omni-flash-preview
- Current output length: 10 seconds (longer durations coming)
- Availability: Google AI Studio, Gemini API
Current limitations:
- Video references up to 3 seconds accepted but not correctly processed
- Audio reference uploads not yet supported
- Character consistency issues across scene changes
- Scene extension not available
The model uses Gemini's multimodal reasoning to synchronize text, graphics, and actions in generated video. Google positions it for conversational editing workflows where users iteratively refine outputs through natural language.
Combined workflow capabilities
Google published three demo applications showing Nano Banana 2 Lite generating initial images that Gemini Omni Flash then animates:
- Anywhere: Generates landmark backgrounds, animates into video clips
- Space Lift: Interior design visualization with cinematic walkthroughs
- Omni Product Studio: E-commerce product video generation
The Interactions API supports up to three sequential edits with maintained session context. Both models include SynthID watermarking for content provenance.
What this means
Google is aggressively pricing Nano Banana 2 Lite to compete in the high-volume image generation market, undercutting alternatives where speed matters more than quality. The $0.034 per 1K image pricing and 4-second latency target prototyping and draft-heavy workflows. However, without independent benchmarks, developers need to validate quality claims against their specific use cases.
The Gemini Omni Flash API opening is more significant. At $0.10 per second, it's the first major video generation model with conversational editing officially available via API, though the 10-second limit and processing bugs (especially around video references) suggest it's genuinely preview-stage. The combined image-to-video pipeline could enable new product categories in e-commerce, marketing, and content creation—if the quality and reliability hold up at scale.
Related Articles
OpenAI Halts Parts of Astra Model Development After It Hit 'Critical' Cybersecurity Threshold
OpenAI disclosed that its in-development Astra model showed cyberattack capabilities strong enough that it cannot rule out a 'Critical' risk classification. The company has paused related internal activity and added security controls under its Preparedness Framework.
Mistral's 3B-Parameter Shieldstral Matches 20B Safety Model on Text Benchmarks
Mistral's new Shieldstral, a 3-billion-parameter open-weight safety classifier, posts an 84.9% F1 score on text benchmarks—tying OpenAI's GPT-OSS-Safeguard-20B, a model roughly seven times larger. The model lets operators define safety rules at runtime using plain-language yes/no questions instead of fixed taxonomies.
Mistral AI Releases Shieldstral-1.0-3B, a 3B-Parameter Policy-Adaptive Safety Classifier
Mistral AI has released Shieldstral-1.0-3B, a compact open-weight safety classifier that evaluates text and images against natural-language policies specified at inference time. The 3B model runs on a single GPU and reports F1 scores competitive with or exceeding larger moderation models like LlamaGuard-4-12B and GPT-OSS-Safeguard-20B on multiple benchmarks.
Black Forest Labs Launches FLUX 3 Video, Claims It Beats Seedance 2.0 on Elo Rankings
Black Forest Labs has made FLUX 3 Video generally available via its API, offering up to 20-second HD/Full HD clips with native audio and lip-sync in 14+ languages. The company claims its internal Elo benchmarks put the model ahead of Seedance 2.0, Gemini Omni Flash, and Minimax H3.
Comments
Loading...