Fresh stories
OpenBMB releases MiniCPM-V 4.7 with 35B-A3B and 1B models
OpenBMB released open MiniCPM-V 4.7 models in 35B-A3B and 1B sizes with mobile deployment support. It reports fivefold faster inference on a 16GB iPad Pro M4 using Lookahead Expert Scheduling.
Artificial Analysis launches AA-3D, a benchmark for interactive scene-building agents
AA-3D evaluates agents on 55 briefs using sandboxed three.js builds and blind human voting. Results compare scene quality across domains and identify models offering the best quality at each cost level.
Cursor joins The X Subscription with shared Grok usage
The X Subscription bundles Cursor, Grok and X Premium with a shared Grok usage pool. Third-party models cost API prices, while existing Cursor subscribers can keep their current plans.
OpenBMB releases MiniCPM-V 4.7 with 35B-A3B and 1B models
OpenBMB released open MiniCPM-V 4.7 models in 35B-A3B and 1B sizes with mobile deployment support. It reports fivefold faster inference on a 16GB iPad Pro M4 using Lookahead Expert Scheduling.
Perplexity Projects commits agent work to S3-backed Git repositories
Perplexity Projects agents commit work from ephemeral sandboxes to S3-backed Git repositories. Short-lived locks and head checks serialize updates, while real checkouts support grep and standard file tools.
Factory makes reusable cloud agent environments generally available
Factory Environments prebuild repositories, tools and dependencies into portable sandbox images. Missions can distribute tasks across sandboxes with consistent starting configurations.
Artificial Analysis launches AA-3D, a benchmark for interactive scene-building agents
AA-3D evaluates agents on 55 briefs using sandboxed three.js builds and blind human voting. Results compare scene quality across domains and identify models offering the best quality at each cost level.
ReadMe launches Restless to repair failed agent API calls
Restless records agent API traffic and adds repair instructions to failed responses. Agents can replay calls through MCP, while missing information can be flagged as Linear issues or suggested documentation changes.
Dan Shipper reports Muse Wonder scored 52% on a codebase rewrite test
Dan Shipper reports Muse Wonder Standard scored 52% versus Sol's 71% after he corrected a faulty test. He tested Standard because Full was unavailable through the API, and other early coding tests produced mixed results.
Top storiesthis week
Addy Osmani outlines hooks to stop coding agents from weakening tests
Addy Osmani describes human-approved completion criteria and hooks that block test edits during implementation. Separate reviews check for weakened assertions, while recurring mistakes become shared hooks or tests.
Claude Code users report unexpectedly fast quota use
Claude Code users report unusually high quota consumption. Boris Cherny says Anthropic made no changes, asks users to share /usage output, and suggests disabling the up-coder plugin to test its impact.
Codex raises Sol’s default context to 500K tokens
GPT-6.1 Sol now defaults to 500K tokens in Codex and ChatGPT Work, up from 272K, to reduce compactions. OpenAI says subscriptions avoid long-context fees, and users can override the default up to 1M tokens.
Zeeg describes a Vercel sandbox harness that stores agent state separately
Zeeg describes a harness using ephemeral Vercel sandboxes with egress controls and separately stored state. Follow-up discussion covers replayable state, credentials, isolation and parallel execution.
Science Made Readable adds MCP access to OpenAlex literature search
Maxime Rivest added an MCP server to Science Made Readable, an open-source semantic search service built on OpenAlex and DenseOn embeddings. Casual access is limited to 10 requests per minute.