Multimodal LLM releases

Moonshot's open coding-focused agentic model built on K2.6, with native vision/video input, forced thinking mode, and stronger long-horizon software-engineering performance.

Native multimodal MiniMax model with a one-million-token context, sparse attention, and agentic coding/cowork positioning.

OpenAI's mid-2026 flagship, headlined by an industry-leading 1.5M-token context window and long-horizon agentic tool use.

The public, guardrailed sibling of Mythos and Anthropic's most capable widely-released model, built for long-horizon agentic work. Launched June 9, 2026 across the Claude API, AWS, and Microsoft Foundry — then pulled three days later under a US government export-control directive barring access by foreign nationals.

Anthropic's most capable model, with strengthened agentic and long-running task performance.

Announced at Google I/O 2026; emphasizes deep multimodal reasoning over a 2M-token context.

xAI's agentic flagship with a 1M-token context and aggressive API pricing.

Google DeepMind's Gemma 4 advanced-reasoning open model for personal computers, part of the April 2026 Gemma 4 family.

Moonshot's open native multimodal agentic model for long-horizon coding, visual interface generation, and autonomous tool orchestration.

Workhorse GPT-5 release with a dedicated Thinking mode; widely deployed across ChatGPT and the API.

Native vision-language MoE supporting 201 languages with a 1M-token context.

Generally available multimodal flagship with native tool use and a 2M-token context.

Introduced genuinely autonomous multi-file coding and stronger computer use.

Open multimodal Kimi model that adds native visual agentic intelligence, instant and thinking modes, and agent-swarm workflows on top of the K2 base.

Open 106B-class vision-language model with native multimodal function calling for visual agents.

Mistral's largest open-weight MoE, aimed at frontier reasoning while remaining self-hostable.

Google's open multimodal model: 128k context, 140+ languages, runs on a single GPU.

Vision-language GLM based on GLM-4.5-Air, covering image, video, document, grounding, and GUI-agent tasks.

xAI's fourth-generation Grok line, preceding the later 4.x API updates already tracked in the catalog.

Baidu's largest ERNIE 4.5 vision-language MoE, supporting text, image, and video inputs with thinking and non-thinking modes.

Updated MIT-licensed Kimi-VL reasoning model with better multimodal reasoning, video understanding, high-resolution perception, and lower thinking-token use.

First Claude 4 Opus model, positioned for long-running agentic and coding work before the 4.x point releases.

Open audio foundation model for audio understanding, generation, speech recognition, audio QA, captioning, and speech conversation.

Efficient MIT-licensed vision-language MoE for OCR, image/video understanding, long documents, and OS-style agent tasks.

Reasoning model released alongside o4-mini with tool use, image reasoning, and stronger agentic problem solving.

API model family focused on coding, instruction following, and one-million-token long-context work.

Meta's flagship open-weight MoE; highest MMLU among open models at release.

Efficient open-weight MoE designed for very long context on modest hardware.

Local omni-modal Qwen model that supports text, image, audio, video, and speech generation in a 7B package.

Reasoning-focused Gemini 2.5 model that made thinking a core part of Google's flagship model line.

Apache-licensed Small update adding vision and a 128K context window to the efficient 24B line.

Anthropic's first hybrid-reasoning Sonnet. Shut down May 11, 2026 as the 4.x line matured.

xAI's third-generation model family, introduced with stronger reasoning, search, and coding modes.

Vision-language Qwen2.5 model for image, document, video, and agentic visual grounding tasks.

Doubao 1.5 Pro update positioned for stronger multimodal, reasoning, and agentic work in Volcano Engine.

Moonshot's multimodal reinforcement-learning reasoning model, reported as matching OpenAI o1 on math, coding, and multimodal reasoning.

Open MiniMax generation with MiniMax-Text-01 and MiniMax-VL-01 long-context models.

Second-generation StepFun foundation model line with larger-scale multimodal and reasoning ambitions.

First Gemini 2.0 release, built for native multimodal input/output, tool use, and agentic product integrations.

General release of OpenAI's o1 reasoning model with stronger deliberative reasoning and multimodal ChatGPT integration.

AWS-native multimodal model with a 300k context; size and architecture undisclosed.

Lower-cost multimodal Nova understanding model for text, image, and video inputs.

Fast, lower-cost Claude 3.5 model for latency-sensitive coding, tool-use, and customer-facing workloads.

First Llama family release with native vision models, alongside smaller edge-oriented 1B and 3B text models.

Open multimodal model family trained for strong image understanding, pointing, and visual grounding.

Mistral's first open multimodal model, adding image understanding to a Mistral text backbone.

Second-generation Grok release with Grok-2 and Grok-2 mini for chat, coding, reasoning, and image-enabled product experiences.

8B vision-language model for local image, multi-image, OCR, and video understanding, with llama.cpp and Ollama support.

Major Sonnet upgrade that became Anthropic's default high-intelligence workhorse for coding, writing, and visual reasoning.

The 2024 omni-modal model that defined a generation of assistants. Deprecated in Feb 2026 and fully retired across ChatGPT on April 3, 2026.

Falcon 2 generation, including text and vision-language 11B models under a permissive TII license.

StepFun's first major vision-language model, released after the Step-1 language model.

Highest-capability Claude 3 model, launched with Sonnet and Haiku and Anthropic's first major vision-capable Claude family.

Gemini generation that introduced production-scale long context, eventually expanding to a two-million-token window.

Zhipu's GLM-4 flagship generation, launched as the successor to ChatGLM3 with stronger tool use and multimodal variants.

Google's first natively multimodal Gemini flagship, since superseded by the 1.5/2/3 lines.

Lower-cost GPT-4 generation with a 128K context window, introduced at OpenAI DevDay.

Baidu's fourth-generation ERNIE flagship, announced with stronger understanding, generation, reasoning, and memory.

Open vision-language assistant and one of the most widely run early local multimodal models.

Second EXAONE generation, improving bilingual Korean-English performance and enterprise deployment options.

The model that brought reliable multi-step reasoning to the mainstream; size never disclosed.

LG AI Research's first EXAONE foundation model generation, introduced as a large multimodal expert AI.