A collection of 15 posts
Kimi K3 is Moonshot AI’s 2.8-trillion-parameter open-weight model, released July 2026. Architecture, specs, pricing, and how to access it.
How Kimi K3 compares to Claude Fable 5, GPT-5.6 Sol, and Opus 4.8 across the Intelligence Index, coding arenas, agentic tasks, and price.
Moonshot's Kimi K2.7 Code and Z.ai's freshly-released GLM 5.2 are both Chinese open-weights coding flagships, both shipped in June 2026, and they trade on opposite axes. K2.7 leads on MCP tool use and pricing; GLM 5.2 leads on 1M context. We pick per workload.
Moonshot's open-weights Kimi K2.7 Code goes head-to-head with Anthropic's Claude Opus 4.8. Architecture, benchmarks (and where they don't exist yet), per-task cost, agentic strength, self-host paths, and a clean per-workload verdict.
Two open-weights heavyweights from China go head-to-head for the agentic-coding throne. K2.7 leads on MCP tool-use depth; V4 leads on raw per-token economics and proven independent benchmarks. We break down cost, agentic strength, self-host paths, and pick a winner per workload.
How Moonshot's open-weight Kimi K2.7 Code stacks up against Claude Opus 4.8, GPT-5.5, and DeepSeek V4 for agentic coding — on price, context, and the benchmarks that exist. K2.7's scores are Moonshot-reported only, so the verdict is subject to change once independent results land.
Moonshot AI's Kimi K2.7 Code — a 1T-parameter open-weight coding model with a 256K context, ~30% fewer thinking tokens than K2.6, and strong MCP tool-use. Benchmarks, pricing, API, and local-deployment guide.
Kimi K2.6 ties GPT-5.5 on SWE-bench Pro at 58.6% — and runs roughly 3x cheaper, with open weights. Where each model wins, with the cost math.
Kimi K2.6 and DeepSeek V4 Pro are the two best open-weights coding models in 2026. K2.6 wins long-horizon agents and swarms; DeepSeek V4 wins on raw price.
Kimi K2.6 ties Opus 4.7 on multilingual SWE-bench but trails by 7 points on Verified — at 1/5th the cost. The honest, benchmark-by-benchmark breakdown.
DeepSeek V4 is out — Pro and Flash tiers, MIT license, 1M context, and pricing that undercuts the frontier by up to 11×. Here's how it stacks up against Qwen3.5, Kimi K2.5, MiniMax M2.7, GPT-5.4, and Claude Opus 4.6.
Kimi-Audio is Moonshot AI's open-source audio foundation model. A single 7B model handles speech recognition, audio understanding, audio generation, and end-to-end voice conversation. The official code targets Linux, so this guide walks Windows users through the setup paths that actually work reliably in 2026 — WSL2, Docker, and the
Quick answer. Kimi-Audio 7B runs on Apple Silicon Macs via MLX-LM for ASR, but speech generation still depends on CUDA-only kernels — pair it with kokoro-tts or parler-tts for Mac TTS. Needs ~20 GB unified RAM, Python 3.11, and HF transformers from main. As of May 2026, no first-party MLX/
In the rapidly evolving field of artificial intelligence, particularly in vision-language models, two notable models have gained attention for their innovative approaches and capabilities: DeepSeek VL2 and Kimi Moonlight 3B. This article aims to provide a detailed comparison of these models, focusing on their architecture, capabilities, performance, and applications. Introduction
Moonshot AI's Moonlight-16B-A3B is a Mixture-of-Experts model with 16B total parameters and ~3B active per token, trained with the Muon optimizer. Released under the MIT license on Hugging Face as moonshotai/Moonlight-16B-A3B-Instruct, it's positioned as Moonshot's compact open-weights model — distinct from the company'