Anthropic: Claude Fable Latest vs Google: Gemini 3 Flash Preview (batch)

Anthropic: Claude Fable Latest vs Google: Gemini 3 Flash Preview (batch): Google: Gemini 3 Flash Preview (batch) has the larger context window (1,048,576 tokens); Google: Gemini 3 Flash Preview (batch) is cheaper on input ($0.25 per 1M tokens); Google: Gemini 3 Flash Preview (batch) supports more input modalities (text, image, file, audio, video). Data via OpenRouter /models, as of Oct 29, 2026.

Source: OpenRouter /models · as of

OpenRouter model catalogue feed

These are provider-catalog specifications, not a task-quality benchmark. Prices describe the source listing and may vary by provider or routing. Unknown values are shown as — when the feed did not supply that value.

Side-by-side comparison

Anthropic: Claude Fable Latest vs Google: Gemini 3 Flash Preview (batch) — data via OpenRouter /models, as of
MetricAnthropic: Claude Fable LatestGoogle: Gemini 3 Flash Preview (batch)
Context (tokens)1,000,0001,048,576
Input $/M$10.00$0.25
Output $/M$50.00$1.50
Input modalitiestext, image, filetext, image, file, audio, video
Output modalitiestexttext
Model id~anthropic/claude-fable-latestgoogle/gemini-3-flash-preview:batch

Frequently asked questions

Which has a larger context window, Anthropic: Claude Fable Latest or Google: Gemini 3 Flash Preview (batch)?

Google: Gemini 3 Flash Preview (batch) has the larger context window at 1,048,576 tokens. (via OpenRouter /models, as of Oct 29, 2026.)

Which is cheaper, Anthropic: Claude Fable Latest or Google: Gemini 3 Flash Preview (batch)?

Google: Gemini 3 Flash Preview (batch) is cheaper on input at $0.25 per 1M tokens. (via OpenRouter /models, as of Oct 29, 2026.)

What input types do Anthropic: Claude Fable Latest and Google: Gemini 3 Flash Preview (batch) support?

Anthropic: Claude Fable Latest accepts text, image, file; Google: Gemini 3 Flash Preview (batch) accepts text, image, file, audio, video.

Model details

← All AI models

Open the prompt composer, then choose a model →

Open the composer and choose a supported target model. The reference catalogue may include models unavailable in the composer.