A collection of 7 posts
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.
JetBrains released Mellum2, a 12B Mixture-of-Experts model that activates just 2.5B parameters per token and ships under Apache 2.0. Here's what it is, where it fits in an AI stack, and how to put it to work.
The 2026 head-to-head: Grok 4.3 vs Claude Opus 4.7 vs Gemini 2.5 Pro on SWE-bench, LiveCodeBench, pricing, real coding workflows, IDE harnesses, and a clear pick-by-job-to-be-done framework.
Gemini CLI vs Claude Code in May 2026: open source vs proprietary, free 1,000 req/day vs $20/mo, SWE-bench scores, install, multimodal workflows, and a clear decision framework.
A neutral 2026 comparison of Claude Code and OpenAI Codex: SWE-bench scores, Terminal-Bench, real pricing, token efficiency, sandboxing, and a clear decision framework for engineering teams.
Cursor Composer 2 vs Claude Sonnet 4.6, with the disambiguation other comparisons skip — Composer is both a feature and a model. Benchmarks, pricing, decision tree, and real workflow patterns for May 2026.