#Coding

AIAnthropic releases Claude Code v2.1.296, which adds a code key to the Claude apps gateway's managed.policies[] and an allow_large option to the Read tool for reading large text files in one call. The release also adds autoCompactWindow for subagents and CLAUDE_CODE_WORKFLOW_SUBAGENT_MODEL, and fixes many bugs in hooks, MCP servers, permission checks and self-hosted runners.

AIGitHub adds two Copilot code review admin controls. Organization owners can bill code reviews from members with a Copilot license to the owning organization instead of member quotas, which requires AI Credits paid usage and allows an optional budget. Owners and repository admins can also restrict review requests to users whose Copilot license comes from their organization or enterprise.

AIOpenAI's developer account announces composer predictions in Codex, which suggest a user's next message based on the conversation and how the user writes. During the beta, the feature is included at no additional cost for eligible Pro users.

AIAnthropic has granted access to Claude Code Projects for every Pro and Max user on the waitlist. Lydia Hallie asks users for feedback on the feature.

AIAnthropic says a Claude project is an ongoing conversation in which Claude coordinates work and runs each task as its own thread in parallel. Documentation is available at

AIaccording to a post from Kilo. The model has 600B parameters, with 27B active per token, and supports a 1M-token context and vision.

AITRAE has merged its TraeCode and TraeWork products into a unified new TRAE with an Agent mode and an IDE mode. In hands-on tests, multiple agents handled planning, design, coding, testing, and fixes within one project, with outputs saved in a shared 'My Artifacts' area. The tests also found that agents working in parallel produced conflicting specifications, so someone had to coordinate them.

AIJetBrains released Mellum2.1, a 12B mixture-of-experts coding model with 2.5B active parameters under Apache 2.0, emphasizing agentic programming. Under high load, its inference throughput in tokens is nearly twice that of Qwen3.5-9B in JetBrains' comparison, and multi-token prediction (MTP) speeds single-request responses by about 1.6x. The model is available on Hugging Face for local or private-infrastructure deployment, with GGUF and vLLM MTP support announced for later.

AIByteDance's AI development brand TRAE has merged its TraeWork and TraeCode clients into a single product with Agent and IDE modes. TraeWork will stop service on October 23, and users who have not upgraded by then will lose access, while chat history, account data and commercial credits migrate automatically.

AITRAE announced on October 9 that it has merged TraeWork and TraeCode into a single platform offering Agent mode and IDE mode with seamless switching between them. The upgraded product covers desktop, web, and mobile, letting users start tasks on a computer, check progress on mobile, and continue development back on desktop.

AIHuawei has opened its DevEco Studio for HarmonyOS PCs to public beta, alongside first public betas of the AI tool DevEco Code and the agent toolkit DevEco CLI. The beta requires HarmonyOS 7.0.0.107 or later, at least 16 GB of memory and 100 GB of storage, and runs on several MateBook models and the MatePad Edge. DevEco Code ships with Zhipu AI's GLM-5.3 and GLM-5.1 models and supports third-party model connections.

AIOpenAI's DevDay 2026 video presents what's new in Codex across the CLI, app, and web. Live demos show moving from delegating individual tasks to having Codex tackle entire problems and goals. The source is a short description only, so no specific features or figures are confirmed.

AIAnthropic is launching OSS Scanner, a service that uses its frontier models to periodically scan opted-in open-source projects for vulnerabilities at no cost. Its reports provide a proof-of-concept, an explanation, and a suggested fix.

AIThariq (@trq212) says the project is built on Claude Managed Agents, with code generated in a cloud sandbox using Sonnet 5.5. He puts each run at about $0.75 and links the GitHub repository for trying it.

AIClaude Code v2.1.295 adds onFailure: "block" for command and HTTP hooks, so a hook that cannot start, times out, or exits unexpectedly blocks the action. The release also adds an optional models list for Claude apps gateway upstreams, plus upstream_request_id in the inference audit event, and fixes a range of MCP, plugin, and terminal issues.

AIOpenAI's Codex 0.162.0 release adds tools for creating and listing managed Git worktrees from trusted local projects when the worktrees feature is enabled. The update also lets users pin tasks in the agent Command Center, copy transcript blocks with /copy, and make URLs clickable in approval headers, questions, and warnings, along with several Linux and Windows sandbox fixes.

AIGitHub has introduced a context-aware classifier that checks candidate secrets in under 2 milliseconds. It could more than double the number of secrets that push protection prevents from entering repository history.

AIStepFun's flagship model Step 5 Preview is free to use in Cline for one week. The quoted background says it scores ahead of Kimi K3 and GLM-5.3 on DeepSWE, positioning it among the strongest open-weights coding models.

AIJetBrains has released Mellum2.1, a 12B mixture-of-experts thinking model with 2.5B active parameters, under Apache 2.0 on Hugging Face. Post-training reinforcement learning in real software repositories raised SWE-bench Verified from 2.0 to 47.0, according to JetBrains' self-reported results. Qwen3.5-9B still leads on SWE-bench Pro, GPQA Diamond and AIME, and GGUF builds start at 7.0 GB for local use.

AIOracle is using ChatGPT Work and Codex to turn specialist knowledge into fast, repeatable workflows across recruiting, engineering, and operations. The source does not provide figures, timelines, or specific results beyond the headline's claim that days of work can take minutes.

AIAnthropic released Claude Haiku 5.5, raising its OSWorld 2.1 score from 15.7% to 72.4% and supporting a 1 million token context window. The article notes Haiku 5.5 still scores 39.2% on Terminal-Bench 4.0 against Sonnet 5.5's 70.6%, and that prompts above 100,000 tokens are priced higher, so migration costs need to be measured on real workloads.

AIStepFun has released Step 5 Preview, its flagship model for agentic work, built on a sparse Mixture-of-Experts architecture with 27B active and 600B total parameters. The source says it performs strongly in software engineering and professional tasks, but the feed supplied only an excerpt, so benchmark details are not available here.

AIGoogle Cloud introduced Gemini agent, a unified AI assistant that acts autonomously, generates code, and completes work across web, mobile, desktop, and third-party apps. It runs jobs on models matched to each task, including Gemini Flash and a flagship frontier model, with Anthropic Claude models also available. Hard spend limits per project let companies enforce budgets and charge AI costs to departments.

AIJetBrains released Mellum2.1, a 12B mixture-of-experts model with 2.5B active parameters under the Apache 2.0 license, built for coding agents. Post-training shifted to reinforcement learning across thousands of environments and millions of sandboxed runs, and the model is available on Hugging Face. The source reports gains over Mellum2 on LiveCodeBench, AIME, GPQA Diamond, BFCL v4, IFEval, and SWE-bench Verified, and says it serves almost twice the tokens of Qwen3.5-9B under heavy load.

Why it matters: The post shows how reinforcement learning in real sandboxed environments changed a compact open model's repository work, with benchmark gains against Mellum2 and two peers.

AIBlock's AI capabilities lead describes using Claude Fable to plan large code migrations and direct smaller models like Opus and Sonnet on individual tasks. He says Block routes frontier and smaller models by task and keeps merges and production deploys behind human dual approval.

Why it matters: Block's engineering lead describes how frontier models orchestrate large migrations and how access, effort levels, and safeguards are managed across an organization.

AINVIDIA reports that fine-tuned Nemotron models reached gold-medal level at both IOI 2026, scoring 535.4 out of 600, and IMO 2026, scoring 30 out of 42. The IOI run was a live, unofficial, unsupervised benchmark, while IMO proofs were graded by official IMO graders. The post also releases checkpoints, datasets, a new 200-problem benchmark, and inference pipelines on Hugging Face and NeMo-Skills.

Why it matters: The post traces how SFT, RL, and a generate-verify-refine loop turned Nemotron into gold-level specialists for IOI and IMO, with the training and inference details shared.

AIAnthropic added build-eval and hillclimb commands to its claude-api skill for designing evaluations and iteratively improving applications against them. The article covers eval design principles, including production-representative tasks, headroom and low variance, and guards against overfitting through train/test splits. Two examples report results: a customer support benchmark where cost fell to under half while accuracy rose, and a claude-api skill eval that rose from 66% to 88%.

Why it matters: The article gives a concrete workflow for designing evals and hillclimbing without overfitting, with two worked cost and performance examples that show the tradeoffs.