Z.ai releases GLM-5.3-Flash and confirms its Ox Alpha identity
Z.ai releases GLM-5.3-Flash with open weights and confirms it was Ox Alpha. It leads DeepSeek-V4-Vision-Exp on six of eight listed coding and agent benchmarks.
EVENT ARCHIVE
Events
Browse published English reports with original sources and content provenance.
AI-generated · awaiting review
OpenAI publishes the first Jalapeño inference benchmarks and a multigeneration roadmap
OpenAI publishes Jalapeño's first InferenceX results on three open models: 1.5–1.9× higher peak performance per watt and 1.7–3.6× lower end-to-end latency.
Qwen releases the open-source Qwen3.8-27B
Qwen releases Qwen3.8-27B, a compact open-source model with visual understanding, adjustable thinking depth and long-horizon agent capabilities.
Z.ai releases GLM-5.3
Z.ai releases GLM-5.3 with stronger long-horizon agent, coding and automated vulnerability-discovery capabilities.
OpenAI previews Cerebras-powered GPT-5.6 Sol Ultrafast
OpenAI opens a limited GPT-5.6 Sol Ultrafast preview to selected API customers, powered by Cerebras and generating up to 750 output tokens per second.
Qwen releases the open Qwen3.8-2.4T-A95B
QwenCloud lists Qwen3.8-2.4T-A95B as the open version of the flagship Qwen3.8 family.
Google releases Gemini 3.7 Flash
Google DeepMind releases Gemini 3.7 Flash, an algorithmic advance over Gemini 3.6 Flash with text, image, audio and video input, text output and up to one million tokens of context. Gemini API documentation lists it as a stable Gemini 3 model.
DeepSeek updates the deepseek-v4-pro API to DeepSeek-V4-Pro-0813
DeepSeek official documentation confirms that deepseek-v4-pro now serves DeepSeek-V4-Pro-0813 under the unchanged API model name, with a one-million-token context window, up to 384K output tokens, and Responses API support.
xAI releases Grok 4.6
xAI releases Grok 4.6 for long-running agents, coding and knowledge work, with a 500K-token context window.
LG AI Research releases the open-weight K-EXAONE 2.0
LG AI Research publishes the K-EXAONE 2.0-750B-A37B model weights and technical report.
Meta releases Muse Spark 1.2 and Muse Code beta
Meta releases the coding-focused Muse Spark 1.2 on August 5, 2026, alongside the Muse Code beta.
DeepSeek V4 Flash enters public API testing with the 0731 release
DeepSeek releases the V4 Flash 0731 API release, with post-training that substantially improves agent capabilities and native support for the Responses API and Codex.
ByteDance officially releases Seedance 2.5
ByteDance officially released Seedance 2.5, which generates up to 30 seconds of audio and video in one pass and accepts as references up to 30 images, 10 video clips, and 10 audio clips.
MiniMax officially releases MiniMax H3
MiniMax officially releases MiniMax H3 on July 31, 2026, and opens the H3-Base weights on August 3.
Google releases Gemini Robotics ER 2
Google DeepMind releases Gemini Robotics ER 2, an embodied-reasoning model for robots with text, image, video and audio input, continuous-video understanding, task orchestration, tool use and multi-robot collaboration.
GPT-5.6 Sol helps optimize OpenAI's production inference stack
OpenAI says GPT-5.6 Sol used Codex to help optimize load balancing, production kernels, speculative decoding and runtime configuration, lowering serving cost and improving generation efficiency.
Google releases Lyria 3.5
Google DeepMind releases the Lyria 3.5 music-generation system. Its model card confirms text prompts as input, music audio and lyric text as output, and distribution through Google Flow Music.
Moonshot AI releases the full Kimi K3 weights
Moonshot AI published the official Kimi K3 repository and full model weights on Hugging Face on July 27, 2026. The 2.8-trillion-parameter multimodal model and its code use the Kimi K3 License.
Qwen3.7 Flash launches on Alibaba Cloud Model Studio
Alibaba Cloud launches qwen3.7-flash and its 2026-07-15 snapshot on Model Studio on July 25, 2026.
Anthropic releases Claude Opus 5
Anthropic releases Claude Opus 5, with upgrades in deep reasoning, long-running agentic coding, vision, professional knowledge work and multi-agent collaboration.