model release

Google releases Nano Banana 2.1 image model: $1.50/$30 per 1M tokens, 66K context

TL;DR

Google's Nano Banana 2.1 (Gemini Nano Banana 2.1) is an image generation and editing model on the Flash tier, listed on OpenRouter at $1.50 input and $30 output per 1M tokens with a 66K context window. It supports 1K, 2K, and 4K output and succeeds Nano Banana 2 and Nano Banana Pro, according to the listing.

3 min read
1

Nano Banana 2.1 at a glance

Context window66K tokens
Input$1.5/1M tokens
Output$7.5/1M tokens

Google's Nano Banana 2.1 (google/gemini-nano-banana-2.1) is an image generation and editing model on the Flash tier, listed on OpenRouter at $1.50 per 1M input tokens and $30 per 1M output tokens with a 66K context window. The listing gives a release date of October 6, 2026.

Key specifications

  • Model: Nano Banana 2.1 (Gemini Nano Banana 2.1), Flash tier
  • Inputs: text and images
  • Outputs: images, with optional text
  • Output resolution: 1K, 2K, and 4K
  • Aspect ratios: extended ratios, set via the image_config API parameter
  • Context window: 66K
  • Pricing: $1.50 input / $30 output per 1M tokens. The provider table also shows a $7.50 per 1M figure under an "Output" column; its exact meaning is not clarified in the listing.
  • Parameter count, training cutoff, benchmark scores: not disclosed

What Google says changed

The listing describes Nano Banana 2.1 as the successor to both Nano Banana 2 and Nano Banana Pro. It claims improvements in product recontextualization, mask- and ink-based editing, and factual accuracy. It also says the model renders photorealistic skin tones, detailed materials, lighting, and coherent backgrounds. No benchmark data or independent evaluations accompany these claims.

Pricing against the Nano Banana family

Using OpenRouter's listed rates for the other Google image models:

Model Context Input / Output per 1M
Nano Banana 2 Lite 66K $0.25 / $30
Nano Banana 2 131K $0.50 / $60
Nano Banana 2.1 66K $1.50 / $30
Nano Banana Pro 131K $2 / $120

Nano Banana 2.1 charges half the output rate of Nano Banana 2 and a quarter of Nano Banana Pro's. Its input rate is three times that of Nano Banana 2. Its context window is half that of both predecessors.

Performance and availability

OpenRouter lists Google AI Studio as the sole provider, with a P50 end-to-end latency of 19.08 seconds. OpenRouter's availability figure for the model over the last 24 hours is 68.62%. The page notes that it can reroute to another healthy provider after an upstream error, but only one provider is currently listed, so that fallback does not apply here. Usage data for apps is not yet available.

The listing does not state whether the model carries SynthID watermarking. Google's smaller Nano Banana 2 Lite is described as doing so.

What this means

Google appears to be splitting the Nano Banana line by cost structure rather than by a single quality ladder. Nano Banana 2.1 pairs a higher input price with a much lower output price than Nano Banana 2 and Pro. Image output tokens dominate generation costs, so per-image costs should fall compared with Nano Banana 2 and Pro, though the listing does not give per-image pricing. Workloads that send many reference images or long prompts will pay more on the input side than they would on Nano Banana 2.

The halved context window (66K versus 131K) is a trade-off for multi-image composition and long-prompt workflows, and teams should test before migrating.

The 68.62% availability and single-provider routing are the main production risks at launch. Until Google publishes benchmarks or third parties run evaluations, the editing and factual-accuracy improvements remain unverified claims.

Related Articles

model release

Thinking Machines Lab releases Inkling 2 Mini: 26B-parameter looped MoE with 3B active and 1M-token context

Thinking Machines Lab has released Inkling 2 Mini, a multimodal Mixture-of-Experts model with about 26B physical total parameters and 3B active per token. It repeats its middle layers three times, supports a 1M-token context, and ships with a DSpark drafter. Hugging Face reports it leads 32 of 48 benchmarks against four comparable models.

model release

Tencent releases Hy4 flagship weights: 770B-parameter MoE, 1,048,576-token context, conditional license

Tencent released downloadable weights for its flagship Hy4 on November 20, with a System Card. The multimodal MoE has a 770-billion-parameter language backbone, 49 billion active parameters per text token and a 1,048,576-token context window, under a license with commercial conditions for model-as-a-service operators.

model release

Gemini 4 Neon scores 61 on Artificial Analysis Intelligence Index, 3rd overall, at $3/$15 per 1M tokens

Google DeepMind has released Gemini 4 Neon, its fourth model in seven weeks. Artificial Analysis measures it at 61 on the Intelligence Index (xhigh), 3rd overall, with token pricing of $3/$15 per 1M input/output tokens and a 1M-token context window.

model release

Google releases Gemini 4 Neon at $3/$15 per 1M tokens with a 1M-token context window

Google released Gemini 4 Neon, available today through the API at $3 per million input tokens and $15 per million output tokens. It is the fourth new Gemini 4 model in four weeks, with a context window of about one million tokens.

Comments

Loading...