Model Updates

New and updated AI models from every major provider.

This Week

Anthropicmodel

Link·Oct 7, 2026

Claude Haiku 5.5

Anthropic's newest small model, built for high-volume, cost-sensitive work. Available in the Claude API and Claude Code.

Best for: Repetitive jobs run at volume: summarizing, classifying, tagging, and pulling data out of documents.
When to use it: You're running the same task hundreds or thousands of times and cost or speed matters more than top-end reasoning.
Who benefits: Teams building their own agents or automations on the Claude API, especially ones that hand smaller steps to a cheaper model.

Perplexitymodel

Link·Oct 7, 2026

pplx-embed-v2-late

For developers building search: two Perplexity models that find matching text and images, including PDF pages, without converting them to text first.

Best for: Building internal search over long or visual documents like slide decks, reports, and scanned PDFs.
When to use it: Your current search misses details in long or image-heavy files because it boils each document down to one summary.
Who benefits: Developers and data teams building retrieval or knowledge-base tools for their company.

SpaceXAImodel

Link·Oct 7, 2026

Grok 4.7 on Microsoft Foundry

SpaceXAI's Grok 4.7 model can now be deployed through Microsoft Foundry, for teams that build on Azure.

Best for: Running Grok 4.7 for coding and long knowledge-work tasks inside an existing Azure setup.
When to use it: Your company's security and billing already run through Azure and you want to test Grok without a separate vendor contract.
Who benefits: IT and engineering teams at Microsoft-standardized companies building internal AI tools.

Perplexitymodel

Link·Oct 6, 2026

pplx-decider v1.1

Perplexity's updated open model for picking the right label or route for a request, at half the price of v1, for developers using the Decisions API.

Best for: Sorting incoming requests or messages into fixed categories, like sending a ticket to billing, support, or sales.
When to use it: You classify inputs with a general chat model today and want a cheaper model built only for that choice.
Who benefits: Developers building triage, routing, or tagging into support and ops workflows.

Googlemodel

Link·Oct 6, 2026

EmbeddingGemma 2

Google's small open model that makes text, images, audio, and video searchable on a phone or laptop, for developers building on-device search.

Best for: Building search that finds a video clip, audio recording, or image from a plain-text query, all on the user's own device.
When to use it: Your app needs to search mixed media without sending files to a server, for privacy or offline use.
Who benefits: App developers adding private or offline search and retrieval to mobile and desktop products.

Googlemodel

Link·Oct 6, 2026

Nano Banana 2.1

Google's updated fast image generation and editing model, now generally available in the Gemini API and replacing Nano Banana 2 for developers.

Best for: Generating and editing images where text inside the image, consistent characters across edits, or very wide formats matter.
When to use it: You build on Nano Banana 2 in the Gemini API and need to move before it shuts down on October 29.
Who benefits: Developers and creative ops teams producing ad variants, banners, and branded images through the API.

Last Week

SpaceXAImodel

Link·Oct 1, 2026

Grok 4.7 on Gemini Enterprise Agent Platform (Preview)

Grok 4.7 is now available in preview inside Google Cloud's Gemini Enterprise Agent Platform (formerly Vertex AI), for Google Cloud customers.

Best for: Using Grok 4.7 inside Google Cloud without a separate SpaceXAI account
When to use it: Your company already builds agents on Google Cloud and wants to test Grok 4.7 next to Gemini
Who benefits: Teams whose IT or data group standardizes on Google Cloud

Googlemodel

Link·Sep 30, 2026

Gemini 4 Argon

Google's new frontier model for long, multi-step work. Today it is limited to a small group of security teams, with paid API customers and AI Ultra subscribers next.

Best for: Long, multi-step work like large codebases, legal and finance analysis, and reports that run far longer than a normal response.
When to use it: Once it reaches the API or AI Ultra, try it on tasks where current Gemini models lose the thread partway through or cut off long outputs.
Who benefits: Developers on paid Gemini API plans and Google AI Ultra subscribers. Most marketing teams will wait for the wider rollout.

OpenAImodel

Link·Sep 29, 2026

GPT-6.1 Sol

An upgrade to GPT-6 Sol that gets close to GPT-6 Astra on coding, computer use, and document work at about a fifth of Astra's price.

Best for: Agent workflows and document-heavy tasks that need strong reasoning but run too often to pay Astra prices every time.
When to use it: Astra is more than the task needs, but GPT-6 Sol keeps missing steps in multi-step workflows or long PDFs.
Who benefits: Developers running agents at volume, and paid ChatGPT Work and Codex users.

Anthropicmodel

Link·Sep 28, 2026

Claude Sonnet 5.5

Anthropic's new everyday model, more than 30% faster than Sonnet 5 and cheaper for most work, available in Claude Code and the API.

Best for: Well-scoped everyday work like fixing bugs, iterating on features, and producing polished documents, slides, and spreadsheets.
When to use it: The task is clear and repeatable and you want speed and lower cost. Switch to Opus 5.5 for long, judgment-heavy work.
Who benefits: Developers and teams running high-volume agent tasks like review, investigation, and drafting on a budget.