Coding Agents

Matt Shumer said a full-access GPT-5.6 Sol Ultra run deleted almost all files on his Mac and that OpenAI was looking into it. Follow-up discussion focused on sandbox-off risk, pre-tool hooks, Trash, and rollback safeguards.

New benchmark posts put GPT-5.6 Sol at or near the top of DeepSWE and several coding/context evals. Cost reports placed Luna on the efficiency frontier, while Amp said replacing Opus with GPT-5.6 cut its average model costs ~50%.

Perplexity added Grok 4.5 as an orchestrator model in Computer for Pro, Max, and Enterprise users. Perplexity reported a WANDR score of 0.328 at $4.76 per trial, while outside security-review and canvas-task tests put it close to GPT-5.6 Sol on cost or token use.

Meta and third-party benchmark posts put Muse Spark 1.1 near frontier coding and agent models at $1.25/M input and $4.25/M output. Results included Vals AI agent tasks, Code Arena Frontend #9, and an AA Coding Agent Index score of 69.

OpenAI launched ChatGPT Work, a Codex- and GPT-5.6-powered ChatGPT agent with a desktop app that can access local files and apps. OpenAI staff said Work and Codex use the same sandboxing with UI changes.

Meta launched Muse Spark 1.1 in Meta AI and the Meta Model API public preview for coding, tool use, computer use, and multimodal reasoning. Early eval posts ranked it highly while system-card threads flagged safety details.

Genspark and OpenClaw added Grok 4.5 after xAI's launch, extending the model into more coding-agent workflows. Follow-up evidence covered AA-Briefcase and Terminal-Bench results, a Composio credential-audit run, and SuperGrok usage-meter reports.

Practitioners described GitHub or folder-based markdown knowledge bases that feed persistent company or personal context to Codex, Claude Code, and Hermes. OpenWiki added codebase and personal brain modes for the same pattern.

OpenAI audited SWE-Bench Pro and found 30% of public tasks were broken. It retracted its earlier recommendation to use the benchmark as a leading coding eval.

Databricks published an internal coding-agent benchmark using tasks from its codebase. OpenAI, Anthropic, and GLM-5.2 models landed on its Pareto frontier, and the company argues teams should optimize cost per task rather than per token.

SpaceXAI launched Grok 4.5 in Cursor and several agent tools with $2/M input and $6/M output pricing. Early evals place it near frontier coding models, with 51% on AutomationBench-AA.

Cognition says SWE-1.7 was trained with RL on a Kimi K2.7 base and now runs in Devin at 1,000 tok/s. It reports 42.3% on FrontierCode at $1.97 per task and released revised grading rules.

VS Code shipped a Copilot update with browser-agent tools for web app validation, plus a preview Agents window for parallel workflows, BYOK model discovery, and cost visibility. Version 1.128 also adds grouped and workspace-less chats in the Agents window.

Posts say Sol, Terra, and Luna are set for a July 9 launch. One report says Sol was added to the Codex codebase as OpenAI’s strongest model for code, research, and documents.

Simon Willison highlighted Jarred Sumner’s Bun rewrite from Zig to Rust as an AI-assisted workflow using trial runs, dynamic planning, adversarial review, and verification. One estimate put API token cost near $165k.

OpenAI says GPT-5.6 Sol, Terra, and Luna will launch publicly Thursday as preview access expands. Testers describe Sol as strong for coding, agents, and computer use; Wafer reports Cerebras serving up to 750 tokens/sec.

Practitioners reported concrete Fable 5 coding outcomes, including sqlite-utils 4.0rc2 for $149.25 and hallucinations in X API and OAuth checks. Failures around tests, finance, production outages, and token-heavy loops kept review systems central.

Engineers debated review depth for AI-written code, from Matt Pocock’s seven-level scale to automated-plus-human review loops. The split was whether pre-patch test failures and deterministic tools add trust, or mock-heavy unit tests just add churn.

Fable users reported cost escalations from $300/day estimates to a single high-effort prompt above $130 and quota-drain complaints on Reddit. Users also reported automatic Opus fallback and safety refusals tied to bio/cyber safeguards.

Threads and papers traced agent reliability failures to harness details including tool schemas, state, retries, logs, and effort-level evals. Examples included Claude Code test loops, MCP server patterns, and OctoTools.

AI SDK added HarnessAgent as a common interface for Pi, Claude, Codex, OpenCode, and other harnesses. Use it to run local or cloud software-factory jobs through official SDKs while subscriptions cover token usage.

Grok Build added speech-to-text dictation for coding agents through /voice or Ctrl+Space. Try it to bring Grok-powered real-time voice input into CLI coding workflows.

Claude Code 2.1.200 changed Manual permission defaults and fixed background-agent crash and recovery paths; 2.1.201 removed mid-conversation Sonnet 5 harness reminders. Update to reduce accidental advances and repeated reminders in stalled sessions.

OpenRouter said four open-weight models now handle real agentic workloads, and a JPMorgan report put Chinese models at about 45% of platform traffic. The shift matters because teams are optimizing for price, hosting, and task fit instead of defaulting to frontier APIs.

Epoch introduced MirrorCode, a benchmark where models reimplement real programs from specs with no internet and hidden held-out tests; the best current score is 56%. The setup matters because it scales inference into multi-day runs and targets software jobs estimated to take humans weeks.

Next.js previewed an agent-focused toolchain with auto-managed AGENTS.md, browser-backed verification, and Skills for cache-component migration and optimization. The release matters because framework guidance, browser introspection, and fix prompts are now packaged directly for coding agents.

OpenAI published usage data showing Codex now generates 99.8% of its internal AI output tokens, with sharp growth in legal, support, recruiting, and finance. The report measures agent adoption as delegated parallel work, not just chat inside engineering.

DeepReinforce released Ornith-1.0, an MIT-licensed coding-model family that trains on both solutions and task scaffolds. The flagship 397B MoE claims 82.4 on SWE-Bench Verified and 77.5 on Terminal-Bench 2.1, pushing open coding models closer to closed frontier systems.

Vals AI launched SkillsBench, a public benchmark for measuring how reusable skills change coding-agent performance, and reported average accuracy rising from 35.5% to 52.5%. The results matter because they suggest some workflows can move to cheaper models when task-specific skills are available.

BrowserCode, Hyper, OpenCode, Together, and other vendors added GLM-5.2 soon after release. That turns the open model into a deployable option across coding, browser automation, and hosted chat.

Builders published Claude Code and Droid setups for GLM-5.2 while Unsloth quantized it for local 256GB machines and Hugging Face opened temporary free inference. Teams can now run the open-weight model across hosted, local, and agent workflows.

Poolside released Apache 2.0 weights for Laguna M.1 and XS.2, its long-horizon coding models, with M.1 shipping at 225B total parameters, 23B active, and 256K context. SGLang and vLLM support on day one lets teams run and fine-tune the models in existing agent stacks immediately.

OpenHands added Agent Client Protocol support to its Agent Canvas, SDK, and Cloud, letting teams run different coding agents through one interface across local, remote, and cloud backends. The release also underpins new OpenHands Index results, so teams can compare harness-plus-model combinations instead of model-only runs.

TryCua brought Cua Driver to Linux, letting Claude Code, Codex, Hermes, and custom agents control real desktop apps via CLI or MCP without taking over the main terminal. The release also adds headless SSH execution and a preview of multi-window Wayland control across supported distros.

Databricks open-sourced Omnigent, a meta-harness that runs Claude Code, Codex, Cursor, Pi, and custom agents in one live session with a collaborative web UI. The release centralizes supervision, cost control, and cross-agent review instead of splitting work across separate tools.

Cursor now lets developers move local agents to the cloud so work can continue after the laptop closes, with mobile as the handoff control surface. The change removes one of the main setup frictions in long-running cloud sessions.

Cursor said it agreed to a $60B all-stock deal with SpaceX, with closing targeted for Q3 and Cursor remaining a wholly owned subsidiary. The deal ties a major coding-agent channel to SpaceX compute and gives Cursor a new strategic owner.

Anthropic paused a same-day policy change that would have moved Claude Agent SDK, claude -p, and third-party SDK apps onto separate monthly credits. Existing subscription-backed workflows continue unchanged for now, but teams should watch for the redesigned billing plan.

Two days after Fable 5 went offline, developers started testing GLM-5.2, GPT-5.5, and multi-model panels against the kinds of one-shot frontend and greenfield builds Fable handled well. The early pattern is that replacements cover much of the work, but Fable still leads on UI taste and first-pass product completion.

OpenRouter, OpenCode, Lovable, Cline, Browser Use Terminal, Nous Portal, and Venice all added Fable 5 within hours of launch. The rollouts put the model into gateways, coding agents, browser agents, and chat clients on day one.

Cohere open-sourced North Mini Code, a 30B-parameter coding MoE with 3B active parameters, 256K context, and Apache 2.0 licensing. OpenCode added it the same day, making the release immediately usable in a coding-agent client.

Cognition introduced FrontierCode, a coding benchmark that grades mergeability and review quality instead of only unit-test passes, and the top model scored 13%. The result matters because it differs from SWE-Bench-style pass rates, and outside researchers are already questioning score variance and reproducibility.

Builders shipped OpenProse workflow files, ghzinga PR tabs, cmux terminal controls, datasette-agent-edit primitives, and an agent-optimized CLI fork. These pieces turn prompt strings into reusable files, panes, and testable edit loops for coding agents.

MIT-linked analysis says AI coding tools sharply raise local code output, but most of the gain disappears by review and release. Teams should watch downstream throughput, since project creation rose without matching demand signals in separate Hugging Face Spaces data.

Uber set a $1,500 monthly limit for each AI coding tool an employee uses, covering products such as Cursor and Claude Code. The cap gives enterprises an early benchmark for coding-agent spend as token costs outgrow typical software-seat budgets.

Two days after Qwen 3.7 Plus launched, Hyper, OpenCode, Kilo, and Vals shipped support or rankings around the 1M-context multimodal model. The rapid pickup shows Alibaba’s new model landing quickly in coding-agent tools and public eval stacks outside its own platform.

Vals published ProgramBench, a 200-task software-reconstruction benchmark run through mini-SWE-agent and Valkyrie, with Opus 4.8 becoming the first model to fully solve two tasks. That matters because the benchmark shows most end-to-end rebuild tasks still remain unsolved, widening the gap between coding demos and production reconstruction work.

A day after MiniMax M3 launched, OpenCode, Hermes Agent, Flowith, Atomic Chat, Kilo Code, Cloudflare AI Gateway, and Vercel AI Gateway shipped support. That breadth shows M3 plugged into agent harnesses and routing layers immediately, not just its own API.

Independent IDEs, gateways, and agent runtimes rolled out Claude Opus 4.8 within hours of launch, including Cursor, Warp, OpenRouter, and Perplexity. That matters because teams can benchmark or swap the model into existing workflows without waiting for connector lag.

DeepSWE launched a coding benchmark built from 113 original tasks across 91 repos and five languages, with GPT-5.5 leading at 70%. The setup is meant to better reflect repo search, multi-file edits, and verification in real agent workflows.