Gemini Developer API — index
Status: navigation page. Provider gemini — https://generativelanguage.googleapis.com/v1beta (key via x-goog-api-key). Cheap probe model: gemini-3.5-flash-lite (the 2.5 family is closed to new keys). Live listing 2026-09-19: 58 ids (v1beta), 22 (v1).
Sources: https://ai.google.dev/gemini-api/docs · https://ai.google.dev/api · sources/gemini/**
Last verified: 2026-09-18.
Platform
| Page | Content |
|---|---|
| Models catalogue | 97 records, Model × Capability / Endpoint / Tool matrices, aliases, lifecycle, live-vs-docs discrepancies (generated) |
| Pricing | every price on the pricing page: text, Live, TTS, transcription, image, video, music, embeddings, tools, agents, tiers |
| Rate limits | usage tiers, spend limits, Priority/Flex, Batch enqueued tokens, Live session limits, 429 semantics, headers (none exposed) |
| Deprecations & changelog | shutdown schedule, API-feature deprecations, 2026 changelog digest |
| Authentication, headers, API versions, regions, terms | API key types, OAuth, ephemeral tokens, alt=sse, v1 vs v1beta (86 vs 47 methods), regions, data-use terms |
| Errors | google.rpc status catalogue, Interactions error codes, finishReason/blockReason, retry matrix, Python/TS handlers |
| SDKs | google-genai 2.24 / @google/genai 2.23 / Go / Java / C# / REST, client options, retries, legacy libraries |
| OpenAI compatibility | /v1beta/openai/* support matrix, reasoning_effort mapping, extra_body fields, live sample |
| Vertex AI vs Gemini API | auth, endpoints, ids, feature parity, data handling |
Capabilities (written by other agents — links may not exist yet)
| Page | Topic |
|---|---|
| generate-content.md | generateContent / streamGenerateContent, contents/parts, generationConfig |
| streaming.md | SSE (alt=sse), chunk shapes |
| structured-outputs.md | responseJsonSchema / responseSchema, enums |
| thinking.md | thinking_level, budgets, thought signatures, summaries |
| context-caching.md | implicit vs explicit caches, TTL, storage pricing |
| safety.md | safetySettings, harm categories, blockReason |
| embeddings.md | gemini-embedding-2 (multimodal, 128–3072 dims), embedding-001 task types |
| files.md | Files API, resumable upload, 48 h TTL, inline vs file data |
| tools/ · ../tools/gemini/index.md | Google Search, Maps, URL context, code execution, computer use, file search, function calling, MCP, tool combination |
| interactions-api.md | Interactions API (stateful, background, agents, store) |
| live-api.md | WebSocket Live API, ephemeral tokens, session resumption, 3.8 Live models |
| image-generation.md | Nano Banana 2 / 2 Lite / Pro |
| video-generation.md | Veo 3.1 (predictLongRunning), Gemini Omni Flash |
| music-generation.md | Lyria 3.5 / 3 / RealTime |
| speech-generation.md | TTS models, voices, streaming TTS |
| transcription.md | gemini-3.5-transcribe (+ live) |
| batch.md | Batch API (batchGenerateContent, asyncBatchEmbedContent), JSONL, 50 % |
| tuning.md | tuning status (retired on the Developer API) |
| file-search.md | File Search stores, indexing pricing |
Machine-readable fragments (this domain)
generated/fragments/models/gemini-models.json · pricing/gemini-pricing.json · rate-limits/gemini-rate-limits.json · deprecations/gemini-deprecations.json · headers/gemini-headers.json · errors/gemini-errors.json · sdks/gemini-sdks.json · endpoints/gemini-models-misc.json. Live inputs: sources/gemini/models-api-raw.json, models-v1-raw.json, models-diff.json (from scripts/discover_gemini_models.py); generator scripts/gen_gemini_models_fragment.py.