Fresh stories
Red-teamers claim Kimi K3 jailbreaks produced cyber and bio outputs
Multiple posts claimed Kimi K3 jailbreaks produced harmful cyber and bio-related outputs. Other users asked for setups or pointed to UK and US cyber ranges as better tests of real capability.
Kimi K3 ranks #5 on Artificial Analysis as engineers dispute coding cost
Kimi K3 posted strong coding results, including rank #5 on Artificial Analysis and #3 on DeepSWE. Engineers disputed whether its lower token price offsets higher token use and slower throughput.
Red-teamers claim Kimi K3 jailbreaks produced cyber and bio outputs
Multiple posts claimed Kimi K3 jailbreaks produced harmful cyber and bio-related outputs. Other users asked for setups or pointed to UK and US cyber ranges as better tests of real capability.
Posts claim GPT-5.6 Sol beats Mythos 5 on UK AISI and CyberGym tasks
Posts citing UK AISI and CyberGym said GPT-5.6 Sol beat Mythos 5 on narrow cyber tasks and The Last Ones. Greg Brockman separately invited defenders to test it on real systems.
Kimi K3 ranks #5 on Artificial Analysis as engineers dispute coding cost
Kimi K3 posted strong coding results, including rank #5 on Artificial Analysis and #3 on DeepSWE. Engineers disputed whether its lower token price offsets higher token use and slower throughput.
Anthropic adds Fable 5 to Max and Team Premium at 50% limits
Starting July 20, Fable 5 is included in Claude Max and Team Premium at 50% of plan limits. Pro and Team Standard shift to credit-based access with a one-time $100 credit.
Anthropic fixes Fable 5 selection outage and issues refunds
Anthropic said it resolved an issue that made Fable 5 unavailable in claude.ai and Claude Code after users reported lost access or credit prompts. ClaudeDevs said affected extra-usage customers would receive refunds plus matching credits.
Developers use Markdown files as long-term memory for agents
Developers described agents writing learned context into linked Markdown pages instead of repeatedly rebuilding it from RAG chunks. Related posts argued Markdown docs and llms.txt can improve agent navigation.
Anthropic releases Claude Managed Agents patterns for long-running agents
Anthropic published Claude Managed Agents examples for long-running production agents. The patterns cover outcome rubrics, self-correction, and sharing sandbox context between agents.
Top storiesthis week
Baseten reports LLM fact writes can retain lift but vanish from answers
Baseten reports that injected facts can retain log-probability lift in weights while disappearing from answers, with prompts recovering 77–80% accuracy. OpenWiki adopted Open Knowledge Format for portable memory files.
Moonshot launches Kimi K3 with 2.8T parameters and 1M context
Moonshot launched Kimi K3 in Kimi products and API with 1M context, native multimodality, KDA/AttnRes, and weights promised by July 27. Benchmarks place it near frontier systems, but testers cite slow serving and usability caveats.
OpenAI traces Codex file deletions to $HOME handling bug
OpenAI says GPT-5.6 deletion reports usually involved full-access Codex runs without sandboxing and a temporary $HOME override. Claude Code 2.1.212 also added loop caps and safety fixes.
LocalLLaMA users report near-6x Qwen 3.6 27B speedups with MTP
A LocalLLaMA benchmark on Qwen 3.6 27B and RTX 6000 PRO reports near-6x speedups from MTP, DFlash, and n-gram drafting. Related tests cover remote prefill, NUMA offload, GPU clock tuning, and llama.cpp Gemma 4 support.
AI21 reports 80.8% on SWE-Bench Pro with model-team agent
AI21 says its pipeline routes exploration, extraction, and patching across model tiers and reaches 80.8% on SWE-Bench Pro at $5.99 per task. Other workflows use critic agents, readable harnesses, Fugu/Nemotron routing, and scheduled Gemini agents.