A collection of 36 posts
Qwen 3.6 is shipping with open weights today. Qwen 3.7-Max was announced May 20 with previews live but no weights yet. A grounded side-by-side.
MiniMax M3.1 Flash went GA on 19 Oct 2026: $0.15/$0.60 per 1M tokens, 1M context, open weights on Hugging Face (minimax-community licence). MiniMax confirmed Space Bunny was an early build. M3.1 Pro: announced, no date. M3 shipped 1 June 2026.
Is Qwen 3.7 released? As of May 2026 it isn't — no weights, API, or benchmarks. Here's what's real, what's only rumored, and what to run today.
SubQ claims to be the first fully subquadratic LLM with a 12M-token context window. Here's what's verified, what isn't, and why the architecture matters.
ZAYA1-8B is an Apache-2.0 MoE reasoning model with 760M active params, pretrained 100% on AMD MI300X GPUs with zero NVIDIA in the loop.
Kimi K2.6 vs DeepSeek V4 vs GLM-5.1 for coding in 2026 — sourced benchmarks, real cost-per-task, self-host and license comparison, plus a clear pick-X-if decision block.
A single tiered table for DeepSeek V4 Pro vs Flash VRAM at every quantization (FP8, FP4+FP8, INT4, Q2) — triangulated across official and community sources.
What OmniCoder 9B is, its lineage and license, vendor-reported benchmarks, the full GGUF quant table, and step-by-step Ollama and llama.cpp setup.
Run Qwen 3.6 locally: 27B dense vs 35B-A3B MoE explained, VRAM tables per quant, and copy-paste Ollama, llama.cpp, vLLM, and MLX commands.
Ant Group's inclusionAI shipped Ring-2.6-1T, a trillion-parameter open-weights reasoning MoE. What it is, the vendor benchmarks, how it stacks up against Kimi K2.6 and DeepSeek V4, and whether you can run it.
AA Index v4.3.2, Oct 26 2026: MiMo-V2.6-Pro 46 (MIT, $0.13/task), DeepSeek V4.1 Flash 1017 45 (MIT), GLM-5.3 45. Step 5 Preview and Kimi K3 44, MiniMax M3.1 Flash 43, Ling 3.1 Flash 41 (MIT), Muse Spark Open 39. Kimi K3.1 Pro (55) weights due Nov 4.