Z.ai released GLM-5.3-Flash, a 320B MoE model with native vision, 1M context, and MIT-licensed weights that matches Claude Opus on coding benchmarks at a tenth of the price. Here’s what held up under independent testing, what it costs, and whether self-hosting makes sense.
Alibaba’s Qwen3.8-Max: A 2.4 Trillion Parameter MoE Model That Actually Keeps Up
Alibaba released Qwen3.8-Max, a 2.4-trillion-parameter MoE model with 1M context, multimodal input, and benchmark results that rival GPT-5.6 and Claude Fable 5. Open weights coming next week.
Kimi K3: Moonshot AI’s 2.8T Open Model Is Here
Moonshot AI just dropped Kimi K3, a 2.8-trillion-parameter open model with a 1M-token context window that benchmarks competitively with Claude Opus 4.8. Here’s what developers need to know about the architecture, benchmarks, and pricing.
ChatGPT Images 2.0: The AI Image Generator That Finally Gets Text Right
ChatGPT Images 2.0 brings reasoning, web research, and accurate text rendering to AI image generation. Here’s what changed, how it compares to Google and Midjourney, and why developers should care.