AI Product Updates: New Features and Releases

ThinkingBox now runs as an OpenEnv environment: it simulates a customer, gives the agent MCP tools over a real database, and judges tasks by what changed in the database rather than the agent's final message, with a 507-task ThinkingBox-Bench.

Binance introduced Binance Intelligence: a free Binance AI assistant, a paid web-based Binance AI Pro, and an Agent OS for developers with MCP access; the rollout is in gray release, with Agent OS already live.

Zernio bundles 16 social platforms, WhatsApp, iMessage, real-number SMS and seven ad networks into one API, with an MCP server and CLI so agents can publish directly.

HashSmash is live, letting AI agents publicly try to break SHA-256, SHA-3, BLAKE3 and Poseidon, with results published and reusable.

Reason Machines released the coding agent Reason 1.0, pitched as fast and lightweight with custom models and subscriptions, strong benchmark scores, and free access.

Agents can post and take jobs from each other, with USDC locked in smart contract escrow and funds released automatically after cryptographic verification of delivery.

Bon Travail is an open-source system in which AI agents find broken code, investigate the problem, and pay human developers to fix it.

ContextQA released Ship, an autonomous quality engineer that tests software, reproduces bugs, investigates failures, and hands the context to Claude Code or Codex.

Ship focuses on continuous background testing, taking a bug report through reproduction, investigation, and handoff to Claude Code or Codex.

ContextQA launched Ship, an "autonomous quality engineer" that tests deployed PRs, reproduces bugs reported in Slack or Linear, investigates the likely cause, and hands fixes to Claude or Codex.

Ship takes a bug report from Slack or Linear through reproduction, investigation, and handoff, positioned as the verification step missing alongside coding agents.

Ship tests deployed PRs, reproduces reported bugs, investigates the cause, and hands fixes to coding agents like Claude or Codex, with a 15-day free trial.