OpenAIAnthropicGooglexAIPerplexity
Nova3 / current state

The state of the frontier, right now.

One read. Refreshed weekly. Drawn from the 40 most material moves across 5 frontier labs in the last 7 days.

By Mark JonesUpdated
Window: Jul 15, 2026 to Jul 22, 2026·Generated Jul 22, 2026, 2:04 PM (CT)
This week's read

Pricing and model references went unstable while agent control moved closer to execution.

The pattern is source surfaces getting thinner on pricing, plan routing, and model references while execution controls move into apps, APIs, and deployed agent stacks. Audit cached pricing and model IDs, then move repeatable work into controls with visible run, admin, and instruction state.

Quiet this window

No material moves from Perplexity in the last 7 days. Tracked, just not active.

Sources informing this read (40)
  1. Jul 22, 2026
    xAIGrok 4.5 is available in the API console for EU users.
  2. Jul 22, 2026
    GoogleGoogle pricing page removed the prior Gemini API pricing content line.
  3. Jul 22, 2026
    GoogleGemini pricing snapshot removed API pricing copy and plan descriptions
  4. Jul 22, 2026
    OpenAIOpenAI modelcard removed guidance naming GPT-5.6 Sol, Terra, and Luna.
  5. Jul 22, 2026
    AnthropicAnthropic removed the pricing-page questionnaire for Team and Enterprise plan recommendations.
  6. Jul 22, 2026
    OpenAIOpenAI model docs removed the previous GPT-5.6 Sol, Terra, and Luna model guidance.
  7. Jul 21, 2026
    AnthropicAnthropic removed the Claude plan recommendation questionnaire from its pricing page.
  8. Jul 21, 2026
    OpenAIOpenAI model docs changed, deleting GPT-5.6 Sol/Terra/Luna selection guidance.
  9. Jul 21, 2026
    AnthropicAnthropic pricing page no longer shows the Claude plan recommendation quiz.
  10. Jul 21, 2026
    GoogleGoogle pricing page removed Gemini API pricing content and Interactions API availability note.
  11. Jul 21, 2026
    GoogleGoogle introduced Gemini 3.6 Flash for everyday tasks in the Gemini app globally.
  12. Jul 20, 2026
    GoogleGoogle pricing diff shows the Gemini API pricing page content was removed.
  13. Jul 20, 2026
    GoogleGoogle pricing page removed Gemini Developer API pricing content from the monitored snapshot
  14. Jul 20, 2026
    OpenAIOpenAI added Record & Replay to Codex on macOS for reusable skills from demonstrated workflows.
  15. Jul 20, 2026
    AnthropicClaude's code execution tool now discloses the per-cell time limit so Claude can budget long-running cells.
  16. Jul 20, 2026
    AnthropicAnthropic pricing removed the self-serve plan recommendation quiz from the monitored page.
  17. Jul 20, 2026
    AnthropicAnthropic is sunsetting the legacy Claude Console Workbench, with access ending on August 17, 2026.
  18. Jul 19, 2026
    AnthropicAnthropic pricing removed the Claude team sizing and plan recommendation quiz content
  19. Jul 19, 2026
    OpenAIChatGPT added help for common questions, easier chat organization, app controls, sharing, and faster iOS photo uploads.
  20. Jul 19, 2026
    OpenAIChatGPT can include text and audio pronunciation guidance for words in more than 60 languages.
  21. Jul 19, 2026
    OpenAIChatGPT now lets users pin chats and projects directly from the sidebar.
  22. Jul 19, 2026
    OpenAIOpenAI improved camera and photo uploads in ChatGPT on iOS, making image adds faster and smoother.
  23. Jul 19, 2026
    OpenAIChatGPT Android added one-off model choice on send, inline mentions, and an updated Plugins menu.
  24. Jul 19, 2026
    AnthropicAnthropic pricing page removed the team sizing and plan recommendation quiz.
  25. Jul 18, 2026
    AnthropicAnthropic model card shows no readable text change in this diff.
  26. Jul 18, 2026
    AnthropicAnthropic pricing page removed the Claude plan recommendation questionnaire content.
  27. Jul 18, 2026
    OpenAIChatGPT scheduled tasks gained a sidebar page for viewing, managing, and editing active tasks.
  28. Jul 18, 2026
    GoogleGemini pricing page removed captured pricing-plan copy and Interactions API banner.
  29. Jul 18, 2026
    AnthropicAnthropic lists Claude Fable 5 as its most capable widely released model.
  30. Jul 18, 2026
    AnthropicAnthropic pricing page removed a five-question plan recommendation flow.
  31. Jul 17, 2026
    AnthropicAnthropic pricing removed the Claude plan quiz recommending Team or Enterprise self-serve.
  32. Jul 17, 2026
    OpenAIOpenAI pricing page removed Business and Enterprise plan copy.
  33. Jul 17, 2026
    AnthropicAnthropic retired Claude Sonnet 4 and Claude Opus 4, and requests to them now return errors.
  34. Jul 17, 2026
    OpenAIOpenAI pricing page removed API pricing table content for listed GPT-5.6 models
  35. Jul 17, 2026
    AnthropicAnthropic removed model card text for Claude Fable 5 and Claude Mythos 5.
  36. Jul 17, 2026
    AnthropicAnthropic removed the monitored model card page’s model overview and comparison content.
  37. Jul 17, 2026
    AnthropicAnthropic pricing page removed team sizing and contract-fit questionnaire content
  38. Jul 17, 2026
    GoogleGoogle pricing snapshot removed Gemini API pricing navigation and availability note.
  39. Jul 17, 2026
    OpenAIOpenAI updated the ChatGPT desktop app for Chat and Work selection, search, Projects, and cross-device Work.
  40. Jul 17, 2026
    OpenAIChatGPT desktop added a ChatGPT and Codex switcher, unified Recents, Projects, and cross-device Cloud Work sync.
FAQ
What is the current state of frontier AI?+

Nova3 publishes a single weekly read synthesizing the most material moves across OpenAI, Anthropic, Google, xAI, and Perplexity in the last 7 days. The full read sits at the top of this page, with every source listed below it.

How often is this updated?+

The pool of frontier moves updates continuously as a Nova3 agent detects new launches. The weekly read is regenerated at least once per 7-day window. The header shows the current window and the last generation timestamp.

Which labs does Nova3 track?+

OpenAI, Anthropic, Google (DeepMind and the Gemini surface), xAI, and Perplexity. Each lab carries an entity link to its Wikidata page so machine readers can resolve the reference unambiguously.

How does Nova3 grade severity?+

Three tiers: critical (changes how something gets built), material (changes who can build it or at what cost), minor (notable but non-structural). A governance layer blocks unsourced claims, so every graded entry carries a primary source link.

Can I get this as raw markdown for an agent?+

Yes. /current-state.md mirrors the same payload as plain markdown for ingestion by LLMs and agentic search clients. /llms.txt and /llms-full.txt expose the broader site corpus.

This is the world we build in. It moves this fast. We move with it.

Bring us your project →

Prefer raw? Read the markdown · See the feed