Skip to content

Documentation Update Tracker

This page tracks all major updates, revisions, and additions to the AI Developer Toolkit Guide, helping you stay current with the latest information and identify what’s changed.

July 2026 - Latest Updates

July 8–11 wave - OpenAI shipped the GPT-5.6 Sol/Terra/Luna family and integrated Codex into the ChatGPT desktop app; Codex CLI reached 0.144.1. Cursor shipped 3.11 and jointly trained Grok 4.5 with SpaceXAI while keeping Composer 2.5 available. Claude Code reached 2.1.207, with account/provider-specific defaults, durable background agents, Routines, dynamic workflows, and current /effort controls.

July 14–19 follow-up - Claude Code reached 2.1.215 (explicit /verify + /code-review skills, the EndConversation tool, /fork to background sessions, screen-reader mode, and Windows PowerShell permission hardening). Codex CLI reached 0.144.6 (stronger dangerous-command detection; GPT-5.6 Sol/Terra/Luna corrected to a 272K working context). Cursor added Slack multi-repo environments and plan-before-start on the 3.11 line.

July 24 wave - Anthropic shipped Claude Opus 5, now the default Opus at Opus 4.8’s $5/$25 pricing and half of Fable 5. It requires Claude Code v2.1.219+, tops the Artificial Analysis Intelligence Index, and carries a May 2026 knowledge cutoff. Two API breaking changes: thinking is on by default, and disabling it is capped at high effort. v2.1.219 also adds category-based fallback — cybersecurity-flagged requests re-run on Opus 4.8, while biology-flagged Opus 5 requests end in a refusal.

Claude Fable 5 - The highest-capability tier remains an explicit selection, not an automatic account default. Its billing settled on July 20, 2026: permanently included on Max and Team Premium at up to 50% of weekly usage limits, with Pro and Team Standard on usage credits.

Verification baseline - Model, pricing, context, effort, and fallback-routing claims were rechecked against Anthropic’s models overview, pricing page, Claude Code model-config docs, and the Artificial Analysis Opus 5 analysis on July 27. Earlier tool-changelog and acquisition-status claims were verified against upstream on July 20.

May 2026 - Latest Updates

Claude Opus 4.8 (May 28) - New flagship, ~4× fewer unflagged code flaws; dynamic workflows + ultracode (research preview, v2.1.154+); /workflows; fast mode 2×/2.5×; lean system prompt default

Claude Code v2.1.144–154 - /resume for background sessions and claude agents --json; /simplify and /code-review are distinct in the current command set

Cursor 3.3–3.5 + Composer 2.5 - PR review, Build in Parallel, full-screen tabs, Composer 2.5 ($0.50/$2.50; fast $3/$15), Jira integration, shared canvases, Automations in the Agents Window

Codex CLI v0.131–0.135 - codex doctor, goals on by default, history search, --profile standardization, and codex remote-control foreground; GPT-5.5 was the default at that historical point

April 2026 - Latest Updates

Cursor 3.0 - Agents Window, Design Mode, Agent Tabs, /worktree, /best-of-n, enterprise features

Claude Code v2.1.84–92 - PowerShell tool (Windows), Bedrock setup wizard, /powerup command, per-model cost breakdown, conditional hooks

ChatGPT desktop (then Codex App) & CLI - Thread search, theming, revamped automations (local/worktree), first-class plugins, sub-agents

Changelog links:

March 2026 - Major Model and Platform Updates

GPT-5.4 released - First general-purpose model with native computer-use, 1M context, default across all Codex/ChatGPT surfaces

Cursor Composer 2 - Frontier coding model (MoE on Kimi K2.5), $0.50/$2.50 per 1M tokens

Codex Security - Application security agent in research preview, 1.2M commits scanned, 792 critical findings

Cursor Automations - Always-on agents with event triggers (Slack, Linear, GitHub, PagerDuty)

Cursor Self-Hosted Cloud Agents - Enterprise: code stays in customer infrastructure

August 2025 - Model Updates

GPT-5 guidance added across model selection, configuration, comparison, and FAQ.

  • Positioning: Great for one-shot app creation from detailed PRDs; experienced developers may still prefer Opus for architecture/planning depth
  • Source: https://cursor.com/blog/gpt-5

July 2025 - Major Release

Version 1.0.0 - Initial comprehensive release

  • Complete guide for Cursor and Claude Code
  • 200+ articles across all sections
  • Enterprise development patterns
  • MCP ecosystem integration
  • Comprehensive appendices

Major Additions:

  • Complete Cursor documentation (100+ articles)
  • Complete Claude Code documentation (100+ articles)
  • Shared workflows section
  • Enterprise development guide
  • MCP ecosystem documentation
  • Comprehensive appendices

Key Features:

  • PRD → Plan → Todo methodology
  • 100 tips for each tool
  • Model comparison guide
  • Pricing calculator
  • Migration checklists
DateSectionChangesImpact
2026-07-27ModelsAdded Claude Opus 5 as the default Opus across model, pricing, routing, and effort guidance; recast Opus 4.8 as the previous tier and cyber-refusal fallback targetHigh
2026-07-27ModelsCorrected Fable 5 plan access (permanently included on Max/Team Premium at 50% of weekly limits since July 20), Gemini 3.1 Pro’s above-200K price tier, and the Opus 4.8 effort defaultMedium
2026-07-20Claude CodeAdded v2.1.208–215 (explicit /verify + /code-review, EndConversation tool, /fork to background sessions, screen-reader mode, permission hardening)Medium
2026-07-20CodexAdded v0.144.2–0.144.6 (stronger dangerous-command detection; GPT-5.6 272K working context)Medium
2026-07-20CursorAdded July 17 Cursor-in-Slack update (plan-before-start, multi-repo environments)Low
2026-07-20ModelsClarified GPT-5.6 ~272K Codex working context vs 1.05M API contextLow
2026-07-11ModelsAdded GPT-5.6 Sol/Terra/Luna, Sonnet 5, and Grok 4.5 across model guides, pricing, and routingHigh
2026-07-11Claude CodeAdded v2.1.185–207 (durable background agents, MCP CLI auth, dynamic-workflow sizing, /doctor)High
2026-07-11CursorAdded 3.8–3.11 (/automate, cloud subagents, iOS, Team MCP marketplaces, side chats)High
2026-07-11CodexAdded v0.142–0.144.1 (Codex in ChatGPT desktop, remote plugins, writes approvals)High
DateSectionChangesImpact
2026-06-10ModelsAdded Claude Fable 5 across model guides, pricing and routingHigh
2026-06-10Claude CodeAdded v2.1.169–170 (--safe-mode, /cd, Claude Fable 5)High
2026-06-08Claude CodeAdded v2.1.156–168 (ultracode trigger keyword, fallback models, .claude/skills auto-load, cloud auto mode)High
2026-06-08CursorAdded 3.6–3.7 (Canvas Design Mode, Cursor SDK, Auto-review Run Mode, Enterprise Organizations)High
2026-06-08CodexConfirmed changelog current at v0.137 (controller pairing, parallel web search)Low
2026-06-08SkillsAdded 11 new official publishers (OpenAI, Google, LangChain, Vercel, Prisma, Convex, Deepgram, Box, Contentful, Contentstack, Base)Medium
DateSectionChangesImpact
2026-05-28Claude CodeAdded Opus 4.8, dynamic workflows + ultracode, v2.1.144–154 changelogHigh
2026-05-28ModelsUpdated defaults to Opus 4.8 / GPT-5.5 / Composer 2.5; refreshed pricingHigh
2026-05-28CursorAdded Cursor 3.4–3.5 and Composer 2.5 changelog entriesHigh
2026-05-28CodexAdded CLI v0.131–0.135 (codex doctor, goals default, history search)Medium
DateSectionChangesImpact
2026-04-05CursorAdded Cursor 3.0 (Agents Window, Design Mode, Agent Tabs)High
2026-04-05Claude CodeAdded v2.1.84–v2.1.92 (PowerShell, Bedrock wizard, /powerup)High
2026-04-05CodexAdded desktop app/CLI updates (thread search, theming, automations; desktop was then branded Codex App)Medium
2026-04-05ModelsDocumented GPT-5.4 as the Codex default in IDE/Cloud surfacesMedium
DateSectionChangesImpact
2026-03-18ModelsGPT-5.4 released as default for Codex/ChatGPT (1M context, computer-use)High
2026-03-18CursorAdded Composer 2 (replacing Composer 1), Automations, Self-Hosted Cloud Agents, Marketplace PluginsHigh
2026-03-12CodexAdded Codex Security research preview, Windows app, CLI 0.117-0.118High
2026-03-12Claude CodeAdded v2.1.81–v2.1.86 (—bare, transcript search, managed-settings)Medium
DateSectionChangesImpact
2026-02-03Claude CodeAdded v2.1.x changelog (v2.1.17-v2.1.30)High
2026-02-03CursorAdded v2.3-v2.4 changelog with Subagents, SkillsHigh
2026-02-03ModelsUpdated GPT-5.1-Codex-Max to GPT-5.2 Codex everywhereMedium
2026-02-03ModelsFixed Opus 4.5 positioning as THE BEST modelHigh
DateSectionChangesImpact
2025-11-24ModelsClaude Opus 4.5 released as default (>80% SWE-Bench)High
2025-11-19ModelsGPT-5.2 Codex for bug fixing and UIMedium
2025-10-29Cursor FeaturesCursor v2.0 with Multi-AgentsHigh
2025-10-29Cursor FeaturesCursor Composer 1 model (4x faster)High
DateSectionChangesImpact
2025-08-07ModelsAdded GPT-5 guidance and referencesHigh
DateSectionChangesImpact
2025-07-17AllInitial releaseMajor
2025-07-16ModelsAdded o3 model detailsMedium
2025-07-15PricingUpdated subscription tiersMedium
2025-07-14MCPAdded 40 server guidesMajor
2025-07-10EnterpriseSecurity compliance sectionHigh

AI Model Changes

Current Models (July 2026):

  • Claude Fable 5 - Highest-capability tier, $10/$50 per 1M, 1M context; thinking always on
  • Claude Opus 5 - The default Opus, $5/$25 per 1M, 1M context; account default on Max, Team Premium, Enterprise pay-as-you-go, Anthropic API, and the managed clouds. Requires Claude Code v2.1.219+
  • Claude Sonnet 5 - Native 1M context; account default on Pro, Team Standard, and Enterprise subscription seats
  • Claude Haiku 4.5 - Cheap, fast tier for subagents and bulk work ($1/$5 per 1M)
  • GPT-5.6 Sol / Terra / Luna - Frontier, balanced, and efficient tiers; 1.05M API context (~272K working context in Codex)
  • Grok 4.5 - Long-running coding and computer work in Cursor ($2/$6 standard)
  • Cursor Composer 2.5 - Frontier coding model, MoE architecture ($0.50/$2.50 per 1M)
  • Gemini 3.1 Pro - Strong multimodal support with 1M context + Deep Think mode

Previous or superseded guidance:

  • GPT-4, GPT-4.1, GPT-4o (legacy relative to current GPT-5.6 guidance; product availability varies)
  • GPT-5.1 / 5.2 and GPT-5.1-Codex-Mini/Max (removed from the Codex picker)
  • GPT-5.3-Codex, GPT-5.4, and GPT-5.5 (superseded by the GPT-5.6 family)
  • Cursor Composer 1 and Composer 2 (older generations; Composer 2.5 is current)
  • Claude Opus 4.8 (superseded by Opus 5 at the same price; not deprecated, still Priority-Tier eligible, and still the fallback target for cybersecurity-flagged Opus 5 and Fable 5 requests)
  • Claude Opus 4.6 and earlier (older than Opus 4.8; v2.1.219 moved the documented managed-cloud defaults to Opus 5, except Foundry, which stays on Sonnet 4.5 and resolves opus to Opus 4.6)
  • Claude Opus 4.1 (deprecated — retires August 5, 2026)
  • Claude 3.x series (legacy); o1-preview, o1-mini (obsolete)

New Features Documented (September 2025):

  • Cursor Browser (Beta) - in-app browser for testing
  • Native Plan Mode (Beta) - AI-assisted task planning
  • Claude Sonnet 4.5 support
  • GPT-5.2 Codex model for bugs and UI

Previous Features:

  • Background Agents (cloud execution)
  • BugBot integration
  • Max Mode enhancements
  • Checkpoints and branching
  • Multi-root workspace support
  • Enhanced @ symbols
  • Memories (beta)

Updated Sections (as of September 2025):

  • Model selection guide (Opus 4.5 was the recommended default at the time; the current Claude Code default depends on account/provider, and managed-cloud defaults differ)
  • Agent mode strategies
  • Context management
  • Team collaboration

Important Changes

Breaking Changes:

  1. Claude Code CLI

    • New installation method via npm
    • Changed config file locations
    • Updated slash command syntax
  2. Cursor

    • New settings structure
    • Changed keybindings for some features
    • Updated extension API
  3. Pricing

    • New tier structure for both tools
    • Changed token limits
    • Updated team pricing
Content TypeSourceLast Verified
PricingOfficial websitesJuly 17, 2025
FeaturesProduct documentationJuly 17, 2025
ModelsAPI documentationJuly 16, 2025
MCP ServersGitHub repositoriesJuly 14, 2025
Tips & TricksCommunity + testingJuly 15, 2025
  1. Official Documentation Review

    • Check vendor release notes
    • Verify API changes
    • Confirm pricing updates
  2. Community Validation

    • Monitor Discord/Slack channels
    • Review GitHub issues
    • Collect user feedback
  3. Hands-on Testing

    • Test new features
    • Verify workflows
    • Update examples
  4. Expert Review

    • Technical accuracy check
    • Best practice validation
    • Enterprise scenario testing

Upcoming Content

August 2025:

  • Local model integration guides
  • Advanced MCP server development
  • Video tutorial supplements
  • Interactive code examples

September 2025:

  • Claude 5 preview documentation
  • New workflow patterns
  • Performance benchmarks
  • Cost optimization strategies
SectionPlanned DateDescription
Video GuidesAugust 2025Embedded video tutorials
Local ModelsAugust 2025Self-hosted AI guides
Advanced MCPSeptember 2025Custom server development
Case StudiesOctober 2025Real-world implementations

Subscribe to our feed:

  • RSS: /en/feed.xml
  • Atom: /en/atom.xml
  • JSON: /en/feed.json

All content includes version tags:

---
version: 1.0.0
lastUpdated: 2026-07-27
tools:
cursor: ">=3.11"
claude_code: ">=2.1.215"
codex_cli: ">=0.144.6"
---
  1. Active - Current best practice
  2. Legacy - Still valid but newer methods exist
  3. Deprecated - Marked for removal
  4. Archived - Historical reference only
Feature/ContentStatusDeprecation DateRemoval Date
GPT-4 examplesDeprecatedJuly 2025October 2025
Old pricing tiersLegacyJuly 2025December 2025
Claude 3.x guidesLegacyJuly 2025January 2026

Documentation Team

Content Creation:

  • Initial guide structure and methodology
  • Tool-specific documentation
  • Enterprise patterns
  • Best practices compilation

Technical Review:

  • Feature verification
  • Code example testing
  • Workflow validation
  • Performance optimization

Community Input:

  • User feedback integration
  • Real-world use cases
  • Tips and tricks collection
  • Problem-solution pairs

Want to contribute? Here’s how:

  1. Report Issues

    • Inaccuracies
    • Outdated information
    • Missing content
    • Broken examples
  2. Submit Updates

    • Fork repository
    • Make changes
    • Submit PR
    • Include verification
  3. Share Experiences

    • Case studies
    • Workflow improvements
    • Performance tips
    • Cost optimizations
MetricTargetCurrentStatus
Accuracy99%99.2%Met
Completeness95%96.5%Met
Clarity90%92.1%Met
Timeliness< 7 days3 daysMet

User Feedback Stats

July 2025:

  • Feedback items received: 247
  • Implemented: 189 (76%)
  • In progress: 34 (14%)
  • Declined: 24 (10%)

Common Themes:

  • More examples requested
  • Video content desired
  • Enterprise scenarios
  • Cost optimization tips

Previous versions available at:

  • /archive/v0.9/ - Beta documentation
  • /archive/v0.8/ - Alpha documentation
  • /archive/drafts/ - Original drafts

When content significantly changes:

  • Migration guide provided
  • Diff highlighting available
  • Reason for change documented
  • Alternative approaches listed
  1. Enable notifications in your account
  2. Select update frequency (immediate/daily/weekly)
  3. Choose sections to monitor
  4. Set importance threshold (all/major/critical)

Report Issues or Suggest Updates

Documentation Issues:

  • GitHub Issues (preferred)
  • Email: docs@[domain]
  • Discord: #documentation

Feature Requests:

  • GitHub Discussions
  • Community Forum
  • Feature Request Form

Urgent Corrections:

  • Security issues: security@[domain]
  • Critical errors: urgent@[domain]

This tracker is updated with every significant documentation change. Last update: July 20, 2026