open-weight models

Z.ai CEO Jie Tang says raw parameter counts no longer predict model quality, pointing to GLM 5.3's benchmark gains that came entirely from reinforcement learning on synthetic long-horizon environments rather than scaling weights. The claim lands alongside a broader reshuffling of agent and legal benchmark leaderboards.

Z.ai's GLM-5.3 ties Kimi K3 for the top spot among open models on the Artificial Analysis Intelligence Index, driven by a major leap in agentic task performance. The company is delaying the open-weight release by about two weeks, citing the model's unusually strong vulnerability-detection capabilities.

Alibaba's Qwen team has released open weights for Qwen3.8, including a 27-billion-parameter multimodal dense model with 262,000 tokens of native context. The models ship under the Apache 2.0 license and are available on Hugging Face and ModelScope.

Alibaba's Qwen team has released FP8-quantized weights for Qwen3.8-27B, a 27-billion-parameter dense vision-language model with native 262,144-token context extensible to 1 million tokens. The model claims gains over its Qwen3.6 and Qwen3.7 predecessors on coding, agentic, and multimodal benchmarks.

Nvidia is reportedly building Nemotron 4, an open-weight model with at least one trillion parameters — double the size of Nemotron 3 Ultra. The company has tripled its cloud spending on in-house training to $28 billion through 2031, with an earliest possible release this fall.

Nvidia's Nemotron 3.5 Lightning, a 31.6B-parameter open-weight model with only 3.6B active parameters, matches OpenAI's gpt-oss-120b on the Artificial Analysis Intelligence Index while delivering the fastest throughput in its class at nearly 670 tokens per second. The model posts especially large gains on agentic benchmarks, beating both gpt-oss-120b and the larger Nemotron 3 Super.

OpenAI has launched ChatGPT Business Premium seats, a new tier priced at $125/month ($100 if billed annually) that offers five times the usage of Standard Business seats and removes the five-hour-per-day limit on advanced features. The move comes as Chinese open-weight models increasingly rival closed-source frontier AI on capability.

Meta released Muse Glimmer, a 30-billion parameter open-weight model under Apache 2.0 license designed to run AI agents locally on consumer hardware. It's an open version of Meta's closed Muse Spark model and supports text, images, and over 100 languages.

Meta has released Muse Glimmer, a 30-billion-parameter open-weight model under Apache 2.0, its first open release since Llama 4 in spring 2025. The launch comes with a Zuckerberg essay defending distillation of rival models and hinting at a future 'dynamic auction' pricing scheme for compute.

A new SaferAI report finds Z.ai's open-weight GLM-5.2 model is only months behind frontier systems like GPT-5.5 and Claude Opus 4.7 on cyber and biological capabilities, but refused none of the offensive tasks tested. Claude Opus 4.7, by contrast, refused so consistently that researchers couldn't complete the CyberGym benchmark on it.

LG AI Research has released K-EXAONE 2.0, a 750-billion-parameter mixture-of-experts language model with 37B active parameters, a 262,144-token context window, and support for 10 languages. The model is open-weighted under Apache 2.0 and claims competitive results against Qwen3.5, GLM-5.1, and DeepSeek-V4 Pro on reasoning, coding, and long-context benchmarks.

Alibaba has released Qwen3.8-Max, a 2.4-trillion-parameter model with 95 billion active parameters per query, designed to run autonomous tasks over multiple days. The company claims it hits 93 on PaperBench and rivals Claude Opus 4.8 and GPT-5.6 Sol on internal benchmarks, with open weights arriving next week.

MiniMax has published MiniMax-H3, a 33-billion-parameter omni-modal generative model capable of producing up to 15 seconds of 2K video with native stereo audio. The model accepts text, image, video, and audio inputs, though its full 2K pipeline depends on a hosted preprocessing component not included in the open-source release.

Thinking Machines Lab has released Inkling Small, an open-weight multimodal mixture-of-experts model with 12B active parameters out of 276B total and a 524K token context window. The model targets reasoning, coding, agentic workflows, and multilingual use cases at $0.58 per 1M input tokens and $1.44 per 1M output tokens.

Moonshot AI released Kimi K3 on July 27, 2026, a 2.8 trillion parameter Mixture-of-Experts model with a 1 million token context window and native multimodal support. AWS has published a deployment guide covering SageMaker HyperPod and Amazon EKS using ml.p6-b300.48xlarge instances with 8 NVIDIA B300 Blackwell Ultra GPUs.

Moonshot AI has released Kimi K3, a 2.8-trillion-parameter open-weight mixture-of-experts model with a 1-million-token context window and native multimodal support. Unsloth has published Dynamic 2.0 quantized versions on Hugging Face, claiming improved accuracy over other quantization methods.

Moonshot AI has released Kimi K3, an open-weight 2.8-trillion-parameter mixture-of-experts model with 104B activated parameters, a 1,048,576-token context window, and native multimodal support. The company describes it as the world's first open 3T-class model, built on a new Kimi Delta Attention architecture.

Poolside has released Laguna S 2.1, a mixture-of-experts coding model with 8 billion active parameters out of 118 billion total, its third coding model release in three months. The company claims it outperforms open-weight models 10 to 20 times its size on agentic coding benchmarks like Terminal-Bench 2.1 and DeepSWE.

Moonshot AI released Kimi K3 on July 16, 2026, featuring 2.7 trillion parameters—the largest open-weight model to date. The company claims K3 performs competitively with Anthropic's Fable 5 while costing $15 per million output tokens compared to Fable's $50.

Amazon Web Services has added three MiniMax models to Amazon Bedrock: M2, M2.1, and M2.5. The newest model, M2.5, uses a mixture-of-experts architecture with 230 billion total parameters and 10 billion active per token, trained specifically for agent-native execution and coding tasks.