Loading page…
Loading page…
Release monitor / Verified records
Every consequential model release, structured.
A source-backed record of AI model launches, developers, availability, licenses, context windows, release types, and disclosed pricing.
Maintained dataset
Records update as sources arrive. Open any row to inspect captured facts, confidence, completeness, and supporting evidence.
Downloadable weights and configuration files for self-hosting; managed access also available through Qwen Cloud
DetailsWhat we captured
Alibaba Cloud’s Qwen team released Qwen3.8-2.4T-A95B, a downloadable open-weight mixture-of-experts model with 2.4 trillion total parameters, 95 billion active parameters, adjustable reasoning effort, and self-hosting support.
Source evidence
Confidence 99% / Completeness 91%
Hugging Face repository; downloadable GGUF files and a 0.9 GB vision projector for local use
DetailsWhat we captured
Dealign.ai published Qwen3.8-27B-CRACK-GGUF, a 27B-parameter, safety-modified derivative of Qwen3.8-27B distributed as seven GGUF quantizations and a vision projector for local llama.cpp inference.
Source evidence
Confidence 92% / Completeness 91%
Available through Nvidia and model repositories including Hugging Face and ModelScope; NeMo Switchyard available on GitHub
DetailsWhat we captured
Nvidia announced Nemotron 3.5 Lightning (30B MoE, 3B active per token), released Aug. 11, with open weights under OpenMDW-1.1 and availability via Nvidia, Hugging Face, and ModelScope; paired with open-source NeMo Switchyard routing library.
Source evidence
Confidence 95% / Completeness 95%
Developers can download and modify the model
DetailsWhat we captured
Meta Platforms released Muse Glimmer, a 30-billion-parameter model available for developers to download and modify.
Source evidence
Confidence 99% / Completeness 87%
Consumer hardware with 24 GB or 32 GB of memory
DetailsWhat we captured
Meta released the 30-billion-parameter multimodal Muse Glimmer model as open weights under Apache 2.0, with quantized configurations intended to run on 24 GB or 32 GB of memory and a 131,072-token context.
Commercial API access
DetailsWhat we captured
Abliteration.ai launched the abliterated-model-large-v2 API, a modified GLM-5.3 designed to reduce refusal behavior for cybersecurity and other sensitive use cases.
Source evidence
Confidence 96% / Completeness 92%
Hosted API globally; downloadable weights restricted to the license's applicable territories
DetailsWhat we captured
MiniMax released H3 on July 31, 2026, offering synchronized stereo-audio video generation through a hosted API and restricted downloadable open weights.
Source evidence
Confidence 98% / Completeness 91%
Available immediately; default on Claude Max and strongest model on Claude Pro
DetailsWhat we captured
Anthropic launched Claude Opus 5, positioning it as a near-frontier model with improved coding, knowledge-work, scientific-research, and agentic capabilities at lower cost than competing frontier models.
Weights published for anyone to download; model can be run locally on user hardware
DetailsWhat we captured
Moonshot introduced Kimi K3 on July 16, 2026 — an open-weights model described as 2.8 trillion parameters, multimodal, with up to a one-million-token context window; its license contains a commercial-use agreement requirement for high-revenue service providers.
General availability following a limited preview
DetailsWhat we captured
OpenAI launched the GPT-5.6 family (Sol, Terra, Luna) for general availability on July 9, 2026, introducing the ultra multi-agent capability and later reducing Luna and Terra prices on July 30, 2026.
Confidence 99% / Completeness 95%
Confidence 92% / Completeness 95%
Confidence 95% / Completeness 87%