Compare models

Context, limits, capabilities, price and use cases side by side — up to 4 models at a time. The URL holds your selection, so you can share it.

Gemini 3.1 Pro PreviewGemini 3.6 Flash Add a modelChat with these 2

AttributeHide identical rows	Gemini 3.1 Pro Previewgemini/gemini-3.1-pro-preview	Gemini 3.6 Flashgemini/gemini-3.6-flash
Specs
Context window	1M	1M
Max output	66K	66K
Input modalities	TextImageAudioVideo	TextImageAudioVideo
Output modalities	Text	Text
Capabilities
Streaming	Supported	Supported
Tool calling	Supported	Supported
Structured output (JSON)	Supported	Supported
Image input	Supported	Supported
Thinking mode	Supported	Supported
Price / 1M tokens
Input	$ 2.00>= 200K$ 4.00	$ 1.50
Output	$ 12.00>= 200K$ 18.00	$ 7.50
Cache read	$ 0.20>= 200K$ 0.40	$ 0.15
Cache write	$ 2.00>= 200K$ 4.00	$ 1.50
What it is good at
Best for	Advanced coding agents Large multimodal corpora	Rapid agentic coding loops Efficient knowledge work Spatial and multimodal reasoning
Published benchmarks
Terminal-Bench 2.0pass_rate	68.5%effort high · Terminus 2(setup not fully disclosed)	—
Terminal-Bench 2.1pass_rate	—	78%Terminus 2(setup not fully disclosed)
BrowseComp (with tools)accuracy	85.9%effort high · tools: browse, python, search(setup not fully disclosed)	—
SWE-Bench Pro (Public)pass_rate	—	58.7%(setup not fully disclosed)
SWE-Bench Verifiedresolved_rate	80.6%effort high(setup not fully disclosed)	—
OSWorld-Verifiedsuccess_rate	—	83%(setup not fully disclosed)
GDPval-AA v2elo	—	1421(setup not fully disclosed)
CharXiv Reasoning (no tools)accuracy	—	85.2%(setup not fully disclosed)
Humanity's Last Examaccuracy	44.4%effort high(setup not fully disclosed)	—
GPQA Diamondaccuracy	94.3%effort high(setup not fully disclosed)	—