When every LLM was released
A continuously updated timeline of large language model releases — when each model shipped from its provider and when it landed on LLM Gateway. Track GPT, Claude, Gemini, Llama, Mistral, DeepSeek and more. Browse the full history by year.
- 304
- models
- 38
- providers
- since
- 2022
Browse by year
Jump to the models released in a given year.
2026
76 modelsAnthropic · Google · Sakana AI · AtlasCloud
Claude Sonnet 5, Gemini 3.1 Flash Lite Image, Grok Imagine Video 1.5
View 2026 releases2025
170 modelsAlibaba · MiniMax · Z.AI · ByteDance
Qwen Image Max 2025-12-30, Gemini 3 Flash (Preview), GPT-5.2 Chat
View 2025 releases2024
48 modelsDeepSeek · xAI · Meta · ElevenLabs
DeepSeek V3, Grok-2 Vision (1212), Grok-2 (1212)
View 2024 releases2023
8 modelsMistral · Anthropic · OpenAI · ElevenLabs
Mixtral 8x7B Instruct, GPT-4 Turbo, GPT-4
View 2023 releases2022
2 modelsOpenAI
GPT-3.5 Turbo
View 2022 releasesLatest releases
The most recently released models, newest first.
Claude Sonnet 5
FlagshipLatestAnthropic · claude-sonnet-5
Claude Sonnet 5 was released by Anthropic on and added to LLM Gateway on .
View model detailsGemini 3.1 Flash Lite Image
FlagshipGoogle · gemini-3.1-flash-lite-image
Gemini 3.1 Flash Lite Image was released by Google on and added to LLM Gateway on .
View model detailsFugu Ultra
Sakana AI · fugu-ultra
Fugu Ultra was released by Sakana AI on and added to LLM Gateway on .
View model detailsKLING v3.0 Turbo
AtlasCloud · kling-v3-0-turbo
KLING v3.0 Turbo was released by AtlasCloud on and added to LLM Gateway on .
View model detailsKLING v3.0
AtlasCloud · kling-v3-0
KLING v3.0 was released by AtlasCloud on and added to LLM Gateway on .
View model detailsGrok Imagine Video 1.5
FlagshipxAI · grok-imagine-video-1-5
Grok Imagine Video 1.5 was released by xAI on and added to LLM Gateway on .
View model detailsGLM-5.2
Z.AI · glm-5.2
GLM-5.2 was released by Z.AI on and added to LLM Gateway on .
View model detailsKimi K2.7 Code Highspeed
FlagshipMoonshot AI · kimi-k2.7-code-highspeed
Kimi K2.7 Code Highspeed was released by Moonshot AI on and added to LLM Gateway on .
View model detailsKimi K2.7 Code
FlagshipMoonshot AI · kimi-k2.7-code
Kimi K2.7 Code was released by Moonshot AI on and added to LLM Gateway on .
View model detailsClaude Fable 5
Anthropic · claude-fable-5
Claude Fable 5 was released by Anthropic on and added to LLM Gateway on .
View model detailsReve Create
Reve · reve-create
Reve Create was released by Reve on and added to LLM Gateway on .
View model detailsQwen3.5 9B
FlagshipAlibaba · qwen3.5-9b
Qwen3.5 9B was released by Alibaba on and added to LLM Gateway on .
View model detailsFAQ
LLM release dates, answered
- How many AI models does LLM Gateway support?
- LLM Gateway currently tracks 304 models from 38 providers across 21 model families, including OpenAI, Anthropic, Google, Meta, Mistral, and DeepSeek. The full list is updated continuously as new models ship.
- What is the newest LLM model on LLM Gateway?
- The most recently released model is Claude Sonnet 5 from Anthropic, released on June 30, 2026. New flagship models usually appear here within 48 hours of their provider launch.
- How quickly are new models added to LLM Gateway?
- New models are typically available on LLM Gateway within 48 hours of their official provider release, so you can switch to the latest model without changing your integration.
- Where do these LLM release dates come from?
- Each entry shows two dates: the official provider release date (when the model maker shipped it) and the date LLM Gateway added support for it. Dates are maintained alongside our model catalog and kept in sync with provider announcements.
- Can I find when a specific model like GPT-5 or Claude was released?
- Yes. Use the search box to filter the timeline by model name, provider, family, or model ID, then read the provider release date for that exact model. Each model also links to a detail page with pricing and capabilities.
Route to any of these models with one API
Switch to the newest model the day it ships — no new SDK, no vendor lock-in. One key for every provider on this timeline.