Turns out the 9router extension I wrote for pi has way more users than I expected. Genuinely thought I was the only one using it.
Turns out the 9router extension I wrote for pi has way more users
More posts
Meta Muse Spark 1.3 is out
. Matches Claude Opus 5 on intelligence benchmarks. Long-context retrieval (MRCR, 1M window): 98.1, best score I've seen. Coding beats GPT-5.6 Sol, roughly ties Claude Fable 5.1. 25% less tokens per task, 20% fewer tool calls. Contributor pricing: $0.10/M tokens. Cheapest model at this tier right now. Meta's been shipping fast on models lately. I think pushing AI into their own social apps gave them a real feedback loop. Better data, faster cycles.
Claude Fable 5.1 dropped today, built on Mythos 5.1
. Coding scores 55.8% on Terminal-Bench, research doubled to 52.6%. Cache is 75% cheaper now, so real costs land 25-45% lower. API prices unchanged.
Meituan shipped LongCat-2.5
. 1.6T params, ~48B active, native multimodal, aimed at long-horizon agents, trained start to finish on Chinese hardware. Images get quantized into multi-level discrete tokens, and that's where Visual Coding and 3D pick up. On the RL side, failed trajectories get turned back into verifiable training tasks. W4A4 inference, 43% faster prefill. Meituan's own evals put it between Claude Opus 4.8 and Opus 5.
SenseTime open-sourced SenseNova Loop-9B-Preview.
It stacks layers instead of parameters. The decoder has 32 layers, but layers 4 through 27 run twice with shared weights, so a forward pass does 56 layers of work while the checkpoint stays at 9B. Their own eval numbers: 71.38 on BIRD-Dev and 13.72 on AutomationBench, both above Qwen3.5-27B. Terminal Bench is still behind.
