Google
US
Gemini 3.8 Live
Google announced this multimodal model release.
MultimodalOfficial source ↗
The frontier, release by release
AI model release timeline
A curated map from ChatGPT’s 2022 debut to today, following major public model launches from OpenAI, Anthropic, xAI, Google, DeepSeek, Qwen and Kimi.
30 Nov 2022 → 18 Sept 2026
Choose the full story or explore one year · official sources
Period Main numbered generation Other major release
RegionType
Company
123Release events
7Active companies
SeptBusiest month
OpenAIMost releases
Swimlane timeline
Hover over any dot for model details. Click a company name to isolate it; click a dot for its source.
Scroll horizontally on smaller screens →
Nov2022
Dec2022
Jan2023
Feb2023
Mar2023
Apr2023
May2023
Jun2023
Jul2023
Aug2023
Sept2023
Oct2023
Nov2023
Dec2023
Jan2024
Feb2024
Mar2024
Apr2024
May2024
Jun2024
Jul2024
Aug2024
Sept2024
Oct2024
Nov2024
Dec2024
Jan2025
Feb2025
Mar2025
Apr2025
May2025
Jun2025
Jul2025
Aug2025
Sept2025
Oct2025
Nov2025
Dec2025
Jan2026
Feb2026
Mar2026
Apr2026
May2026
Jun2026
Jul2026
Aug2026
Sept2026
OpenAI33 releases
Anthropic18 releases
xAI13 releases
Google19 releases
DeepSeek12 releases
Qwen20 releases
Kimi8 releases
Release cadence
Number of release events per calendar month for the current filter.
Release index
Newest first. Each item links to an official announcement or official model card.
DeepSeek
China
DeepSeek V4.1 Flash
DeepSeek announced this multimodal model release.
MultimodalOfficial source ↗
OpenAI
US
GPT-Live-1
OpenAI announced this multimodal model release.
MultimodalOfficial source ↗
OpenAI
US
ChatGPT Images 2.5
OpenAI announced this multimodal model release.
MultimodalOfficial source ↗
OpenAI
Main generationUS
GPT-6 Astra
OpenAI’s sixth-generation flagship model release.
GeneralOfficial source ↗
Qwen
China
Qwen3.8 Flash-Next
Qwen announced this general model release.
GeneralOfficial source ↗
DeepSeek
China
DeepSeek V4 Flash Vision Exp
DeepSeek announced this multimodal model release.
MultimodalOfficial source ↗
Google
US
Gemini 3.7 Flash
Google announced this general model release.
GeneralOfficial source ↗
DeepSeek
China
DeepSeek V4 Pro
DeepSeek announced this general model release.
GeneralOfficial source ↗
xAI
US
Grok 4.6
xAI announced this general model release.
GeneralOfficial source ↗
Qwen
Main generationChina
Qwen3.8
Qwen’s largest open generation, spanning a 2.4T mixture-of-experts flagship and smaller models.
GeneralOfficial source ↗
xAI
US
Grok Imagine Image 2
xAI announced this multimodal model release.
MultimodalOfficial source ↗
Anthropic
Main generationUS
Claude Opus 5
Anthropic’s fifth-generation flagship Opus model for its most demanding reasoning and agentic work.
GeneralOfficial source ↗
Google
US
Gemini 3.6 Flash family
Google announced this general model release.
GeneralOfficial source ↗
xAI
US
Grok 4.5
xAI announced this general model release.
GeneralOfficial source ↗
Kimi
Main generationChina
Kimi K3
Moonshot AI’s new frontier generation: natively multimodal with a one-million-token context.
GeneralOfficial source ↗
OpenAI
US
GPT-5.6 Sol, Terra + Luna
OpenAI announced this general model release.
GeneralOfficial source ↗
OpenAI
US
GPT-Live
OpenAI announced this multimodal model release.
MultimodalOfficial source ↗
Anthropic
US
Claude Sonnet 5
Anthropic announced this general model release.
GeneralOfficial source ↗
OpenAI
US
GPT-5.6 Sol preview
OpenAI announced this general model release.
GeneralOfficial source ↗
Google
US
Gemini 3.5 Flash Computer Use
Google announced this specialized model release.
SpecializedOfficial source ↗
xAI
US
Grok Imagine Video 1.5
xAI announced this multimodal model release.
MultimodalOfficial source ↗
Anthropic
US
Claude Fable 5 + Mythos 5
Anthropic announced this general model release.
GeneralOfficial source ↗
xAI
US
Grok Imagine 1.5
xAI announced this multimodal model release.
MultimodalOfficial source ↗
Anthropic
US
Claude Opus 4.8
Anthropic announced this general model release.
GeneralOfficial source ↗
Google
US
Gemini 3.5
Google announced this general model release.
GeneralOfficial source ↗
OpenAI
US
GPT-5.5 Instant
OpenAI announced this general model release.
GeneralOfficial source ↗
DeepSeek
Main generationChina
DeepSeek V4 Preview
DeepSeek’s fourth generation introduced Pro and Flash models with one-million-token context.
GeneralOfficial source ↗
OpenAI
US
GPT-5.5
OpenAI announced this general model release.
GeneralOfficial source ↗
OpenAI
US
ChatGPT Images 2.0
OpenAI announced this multimodal model release.
MultimodalOfficial source ↗
Kimi
China
Kimi K2.6
Kimi announced this general model release.
GeneralOfficial source ↗
Anthropic
US
Claude Opus 4.7
Anthropic announced this general model release.
GeneralOfficial source ↗
OpenAI
US
GPT-Rosalind
OpenAI announced this specialized model release.
SpecializedOfficial source ↗
Qwen
China
Qwen3.6
Qwen announced this general model release.
GeneralOfficial source ↗
Google
US
Gemini 3.1 Flash Live
Google announced this multimodal model release.
MultimodalOfficial source ↗
OpenAI
US
GPT-5.4 mini + nano
OpenAI announced this general model release.
GeneralOfficial source ↗
OpenAI
US
GPT-5.4
OpenAI announced this general model release.
GeneralOfficial source ↗
OpenAI
US
GPT-5.3 Instant
OpenAI announced this general model release.
GeneralOfficial source ↗
Google
US
Gemini 3.1 Flash-Lite
Google announced this general model release.
GeneralOfficial source ↗
Google
US
Gemini 3.1 Pro
Google announced this general model release.
GeneralOfficial source ↗
Google
US
Lyria 3
Google announced this specialized model release.
SpecializedOfficial source ↗
Anthropic
US
Claude Sonnet 4.6
Anthropic announced this general model release.
GeneralOfficial source ↗
Qwen
China
Qwen3.5
Qwen announced this general model release.
GeneralOfficial source ↗
OpenAI
US
GPT-5.3-Codex-Spark
OpenAI announced this code model release.
OpenAI
US
GPT-5.3-Codex
OpenAI announced this code model release.
Anthropic
US
Claude Opus 4.6
Anthropic announced this general model release.
GeneralOfficial source ↗
Qwen
China
Qwen3-Coder-Next
Qwen announced this code model release.
Qwen
China
Qwen3-ASR
Qwen announced this specialized model release.
SpecializedOfficial source ↗
Kimi
China
Kimi K2.5
Kimi announced this general model release.
GeneralOfficial source ↗
Qwen
China
Qwen3-TTS
Qwen announced this specialized model release.
SpecializedOfficial source ↗
Qwen
China
Qwen-Image-2512
Qwen announced this multimodal model release.
MultimodalOfficial source ↗
OpenAI
US
GPT-5.2-Codex
OpenAI announced this code model release.
Google
US
Gemini 3 Flash
Google announced this general model release.
GeneralOfficial source ↗
OpenAI
US
GPT-Image-1.5
OpenAI announced this multimodal model release.
MultimodalOfficial source ↗
OpenAI
US
GPT-5.2
OpenAI announced this general model release.
GeneralOfficial source ↗
Google
US
Gemini 3 Deep Think
Google announced this general model release.
GeneralOfficial source ↗
DeepSeek
China
DeepSeek V3.2 + Speciale
DeepSeek announced this general model release.
GeneralOfficial source ↗
Anthropic
US
Claude Opus 4.5
Anthropic announced this general model release.
GeneralOfficial source ↗
OpenAI
US
GPT-5.1-Codex-Max
OpenAI announced this code model release.
xAI
US
Grok 4.1 Fast
xAI announced this general model release.
GeneralOfficial source ↗
Google
Main generationUS
Gemini 3
Google’s third-generation flagship, built for advanced reasoning, multimodality and agents.
GeneralOfficial source ↗
xAI
US
Grok 4.1
xAI announced this general model release.
GeneralOfficial source ↗
OpenAI
US
GPT-5.1
OpenAI announced this general model release.
GeneralOfficial source ↗
Kimi
China
Kimi K2 Thinking
Kimi announced this general model release.
GeneralOfficial source ↗
Anthropic
US
Claude Haiku 4.5
Anthropic announced this general model release.
GeneralOfficial source ↗
OpenAI
US
Sora 2
OpenAI announced this multimodal model release.
MultimodalOfficial source ↗
Anthropic
US
Claude Sonnet 4.5
Anthropic announced this general model release.
GeneralOfficial source ↗
DeepSeek
China
DeepSeek V3.2-Exp
DeepSeek announced this general model release.
GeneralOfficial source ↗
DeepSeek
China
DeepSeek V3.1 Terminus
DeepSeek announced this general model release.
GeneralOfficial source ↗
Qwen
China
Qwen3-VL 235B
Qwen announced this multimodal model release.
MultimodalOfficial source ↗
xAI
US
Grok 4 Fast
xAI announced this general model release.
GeneralOfficial source ↗
OpenAI
US
GPT-5-Codex
A GPT-5 variant optimized for agentic coding in Codex.
Qwen
China
Qwen3-Next
A more efficient Qwen architecture designed for long context and high throughput.
GeneralOfficial source ↗
Kimi
China
Kimi K2 Instruct 0905
A K2 update with stronger coding, agentic and frontend capabilities.
GeneralOfficial source ↗
DeepSeek
China
DeepSeek V3.1
A hybrid model supporting both thinking and non-thinking response modes.
GeneralOfficial source ↗
OpenAI
Main generationUS
GPT-5
OpenAI’s unified flagship generation, combining fast responses with deeper reasoning.
GeneralOfficial source ↗
Anthropic
US
Claude Opus 4.1
An Opus update focused on agentic tasks, real-world coding and reasoning.
GeneralOfficial source ↗
Qwen
China
Qwen-Image
Qwen’s native image generation foundation model.
MultimodalOfficial source ↗
Qwen
China
Qwen3-Coder
A code-focused Qwen3 model designed for agentic software engineering.
Kimi
Main generationChina
Kimi K2
Moonshot AI’s open mixture-of-experts model for coding and agentic tool use.
GeneralOfficial source ↗
xAI
Main generationUS
Grok 4
xAI’s fourth-generation flagship reasoning model.
GeneralOfficial source ↗
DeepSeek
China
DeepSeek R1-0528
A major R1 reasoning update with improved depth and benchmark performance.
GeneralOfficial source ↗
Anthropic
Main generationUS
Claude 4: Opus + Sonnet
Anthropic’s fourth generation, built for advanced reasoning, coding and long-running agents.
GeneralOfficial source ↗
Qwen
Main generationChina
Qwen3
Qwen’s hybrid thinking generation spanning dense and mixture-of-experts open models.
GeneralOfficial source ↗
OpenAI
US
OpenAI o3 + o4-mini
Reasoning models capable of combining tools, vision and deliberate analysis.
GeneralOfficial source ↗
OpenAI
US
GPT-4.1 family
GPT-4.1, mini and nano improved coding, instruction following and long-context use.
GeneralOfficial source ↗
Kimi
China
Kimi-VL
Moonshot AI’s open vision-language model for visual reasoning and document understanding.
MultimodalOfficial source ↗
Google
Major releaseUS
Gemini 2.5 Pro
Google’s thinking-model generation, reasoning before responding on complex tasks.
GeneralOfficial source ↗
OpenAI
US
GPT-4.5
A research preview focused on broader knowledge, natural conversation and reduced hallucination.
GeneralOfficial source ↗
Anthropic
US
Claude 3.7 Sonnet
Anthropic’s first hybrid reasoning model, combining near-instant and extended thinking modes.
GeneralOfficial source ↗
xAI
Main generationUS
Grok 3
xAI’s third-generation flagship, trained with substantially expanded compute.
GeneralOfficial source ↗
Google
US
Gemini 2.0 family
Gemini 2.0 Flash reached general availability with new Pro and Flash-Lite variants.
MultimodalOfficial source ↗
OpenAI
US
OpenAI o3-mini
A compact reasoning model optimized for science, mathematics and coding.
GeneralOfficial source ↗
Qwen
China
Qwen2.5-Max
Alibaba’s large mixture-of-experts Qwen flagship.
GeneralOfficial source ↗
DeepSeek
Major releaseChina
DeepSeek R1
An open reasoning model that drew global attention for capability and training efficiency.
GeneralOfficial source ↗
Kimi
Main generationChina
Kimi K1.5
Moonshot AI’s multimodal reasoning model, trained with large-scale reinforcement learning.
MultimodalOfficial source ↗
DeepSeek
Main generationChina
DeepSeek V3
A large mixture-of-experts model that established DeepSeek as a frontier competitor.
GeneralOfficial source ↗
Google
Main generationUS
Gemini 2.0 Flash
Google’s agentic-era Gemini generation with native tool use and multimodal output.
MultimodalOfficial source ↗
Qwen
China
QwQ-32B-Preview
Qwen’s early open reasoning model preview.
GeneralOfficial source ↗
Anthropic
US
Claude 3.5 Sonnet (new) + Haiku
An upgraded Sonnet and a new fast Haiku model, alongside computer-use capability.
GeneralOfficial source ↗
Qwen
China
Qwen2.5
A major Qwen refresh with improved knowledge, coding, mathematics and instruction following.
GeneralOfficial source ↗
OpenAI
Major releaseUS
OpenAI o1 preview
OpenAI’s first reasoning-series model, trained to spend more time thinking before answering.
GeneralOfficial source ↗
DeepSeek
Main generationChina
DeepSeek V2.5
A combined general and coding model that unified DeepSeek-V2-Chat and Coder-V2 capabilities.
GeneralOfficial source ↗
Qwen
China
Qwen2-VL
A vision-language model family for image, video and multilingual visual understanding.
MultimodalOfficial source ↗
xAI
Main generationUS
Grok-2
xAI’s second generation added stronger reasoning and image generation in the Grok app.
MultimodalOfficial source ↗
Anthropic
US
Claude 3.5 Sonnet
The first Claude 3.5 model, improving intelligence, speed and coding performance.
GeneralOfficial source ↗
Qwen
Main generationChina
Qwen2
A multilingual open model generation covering 0.5B through 72B parameters.
GeneralOfficial source ↗
Google
US
Gemini 1.5 Flash
Google’s faster, lower-cost member of the Gemini 1.5 family.
MultimodalOfficial source ↗
OpenAI
Major releaseUS
GPT-4o
A real-time omni model designed to reason across text, vision and audio.
MultimodalOfficial source ↗
xAI
US
Grok-1.5
A stronger reasoning model with a 128K context window.
GeneralOfficial source ↗
xAI
Main generationUS
Grok-1 open weights
xAI released Grok-1’s base model weights and architecture under Apache 2.0.
GeneralOfficial source ↗
Anthropic
Main generationUS
Claude 3 family
Claude 3 launched Haiku, Sonnet and Opus with native vision and a new capability tier.
MultimodalOfficial source ↗
Google
US
Gemini 1.5 Pro
Introduced a mixture-of-experts architecture and breakthrough long-context capability.
MultimodalOfficial source ↗
Qwen
Main generationChina
Qwen1.5
A broad open-weight refresh spanning small local models through 72B parameters.
GeneralOfficial source ↗
Google
Main generationUS
Gemini 1.0
Google launched its natively multimodal Gemini generation in Ultra, Pro and Nano sizes.
MultimodalOfficial source ↗
OpenAI
US
GPT-4 Turbo
A faster GPT-4 with a 128K context window and updated knowledge.
GeneralOfficial source ↗
OpenAI
US
GPT-4V in ChatGPT
ChatGPT gained native image understanding alongside voice interaction.
MultimodalOfficial source ↗
Qwen
Major releaseChina
Qwen-7B
The first open release in Alibaba’s Qwen model family.
GeneralOfficial source ↗
Anthropic
Main generationUS
Claude 2
A new Claude generation with stronger reasoning and a substantially longer context window.
GeneralOfficial source ↗
Google
US
PaLM 2
Google’s multilingual reasoning model family, launched across products and Cloud.
GeneralOfficial source ↗
OpenAI
Main generationUS
GPT-4
OpenAI’s first broadly released multimodal GPT generation, accepting images and text.
GeneralOfficial source ↗
Anthropic
Main generationUS
Claude
Anthropic introduced Claude and Claude Instant for business and consumer use.
GeneralOfficial source ↗
OpenAI
Main generationUS
ChatGPT research preview
The public launch that brought conversational generative AI into the mainstream.
GeneralOfficial source ↗
Scope: major public, named model launches. Silent checkpoints, routine API snapshots, pricing-only changes and cloud availability announcements are excluded. A launch containing several variants is counted as one event.30 Nov 2022–18 Sep 2026 · Dates use the publisher’s announcement date.
