Google AI Studio (Gemini API)

permanent free · model_vendor · verification: verified (85%, mixed, 2026-07-17) · setup guide →

The record

Base URLs	https://generativelanguage.googleapis.com/v1beta native https://generativelanguage.googleapis.com/v1beta/openai openai-compatible
API key required	yes
Credit card	no
Phone verification	unknown
May train on prompts	yes
Quota	free_tier [rpm, rpd, tpm] “Gemini 2.5 Flash: 250,000 tokens/minute, 20 requests/day, 5 requests/minute; Gemini 3.1 Flash-Lite: 250,000 tokens/minute, 500 requests/day, 15 requests/minute; Gemma 3 27B Instruct: 15,000 tokens/minute, 14,400 requests/day, 30 requests/minute (per model)”
Signup friction	3 steps · ~5m to first key
Env var	GEMINI_API_KEY
Notes	OpenAI-compatible base URL confirmed from official docs: https://generativelanguage.googleapis.com/v1beta/openai/ (streaming, function calling, and image input all documented as supported). Regional caveat conflict between repo sources: mnfst says free tier unavailable in EU/UK/Switzerland; cheahjs says data is used for training only when used outside UK/CH/EEA/EU — unresolved, treat availability in those regions as unknown.

model	capabilities	context	free	limits
gemini-2.5-flash-lite ⚠ deprecated	chat, vision, tool_calling	?	yes	10 rpm · 20 rpd · 250000 tpm
gemini-2.5-flash	chat, vision, audio_in, tool_calling, reasoning	1,048,576	yes	5 rpm · 20 rpd · 250000 tpm
gemini-3.5-flash	chat, vision, audio_in, tool_calling, reasoning	1,048,576	yes	5 rpm · 20 rpd · 250000 tpm
gemma-3-27b-it	chat	?	yes	30 rpm · 14400 rpd · 15000 tpm