Fresh stories
Study reports Claude Code and Codex memory can store prompt-injection rules
A reported study found Claude Code and Codex resisted credential exfiltration but could store rules that later triggered unauthorized tools. Related memory work warns personal memories can raise hallucinations and RAM pressure.
Kimi K3 benchmarks last at 53/67 in AlphaSignal repair harness
AlphaSignal's repair harness put Kimi K3 last at 53/67, while other tests ranked it high on DeepSWE, Vibe Code Bench, legal work, cyber, and CUDA kernels. Cost often beat Fable, but speed lagged.
Code Arena users report Kaleb shows Qwen-like token quirks
Multiple testers reported Kaleb and torenia-alpha appearing in Code Arena/LMArena, and several pointed to Qwen-like token quirks in Kaleb. Follow-up tests described strong 3D output and a late-2025 or early-2026 cutoff, but the model identity remains unconfirmed.
Study reports Claude Code and Codex memory can store prompt-injection rules
A reported study found Claude Code and Codex resisted credential exfiltration but could store rules that later triggered unauthorized tools. Related memory work warns personal memories can raise hallucinations and RAM pressure.
ChatGPT Work supports Plus, Pro, Business, and Enterprise on web and mobile
OpenAI staff said ChatGPT Work is available on web and mobile for Plus, Pro, Business, and Enterprise. Early users showed sites, email, documents, sheets, and browser tasks, while asking for memory and clearer docs.
Users claim DeepSeek V4 routes hard API prompts through Claude Fable 5
OpenCode users reported complex DeepSeek V4 API prompts returning Fable-like outputs and CoT style, while simpler prompts did not. The evidence is community-led and disputed, so the claim remains unverified.
Kimi K3 benchmarks last at 53/67 in AlphaSignal repair harness
AlphaSignal's repair harness put Kimi K3 last at 53/67, while other tests ranked it high on DeepSWE, Vibe Code Bench, legal work, cyber, and CUDA kernels. Cost often beat Fable, but speed lagged.
Developers report GPT-5.6 Sol and Fable overengineer small coding tasks
Slate, LangChain, and AI SDK add graph control for agent runtimes
Posts say Anthropic makes Fable 5 permanent on Claude Max and Team Premium at 50% limits
Code Arena users report Kaleb shows Qwen-like token quirks
Top storiesthis week
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.
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.
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 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.