
Kimi K3.1 Pro: Open Weights, Near-Frontier Coding
Moonshot's Kimi K3.1 Pro scores 65.7% on Terminal-Bench 4.0, 0.7 points behind Claude Opus 5.5, at half the price. Full open weights are promised by 4 November.
Tag archive

Moonshot's Kimi K3.1 Pro scores 65.7% on Terminal-Bench 4.0, 0.7 points behind Claude Opus 5.5, at half the price. Full open weights are promised by 4 November.
On October 15, Meta released the weights of Muse Spark Open , an 816-billion-parameter model it...
Am 15. Oktober hat Meta die Gewichte von Muse Spark Open veröffentlicht, einem Modell mit 816...
Innerhalb von vier Tagen wurden drei Open-Weight-Sprachmodelle angekündigt. Am 3. Oktober 2026, dem...
Three open-weight language models were announced in four days. On October 3, 2026 — the Day of...

Qwen3.8-Flash-Next packs 125B parameters but activates only 6B per token. How it compares to Qwen3.8-27B, and when the sparse MoE wins.

Qwen3-Coder-480B wins on licence and reported SWE-bench Verified score. GLM-5.2 wins on agentic benchmarks. Here is which one to actually download.
Fine-tuning open models in 2026: when it beats RAG, and how eCorpIT delivers it Summary....