Two AI hosts break down a week where three frontier labs shipped new models within days of each other. OpenAI released the GPT-5.6 family and the ChatGPT Work agent, xAI put out Grok 4.5 at $2 per million input tokens, and Meta charged for access to one of its own frontier models for the first time with Muse Spark 1.1. The conversation follows the pricing race down to the hardware underneath it, from SK Hynix's $26.5 billion IPO to the billion-dollar compute deals keeping everything running. Companion to this week's AI Vaults newsletter, Three Frontier Models in Seven Days.

What's Covered

* OpenAI's GPT-5.6 family (Sol, Terra, and Luna) and the ChatGPT Work agent

* Grok 4.5 pricing and xAI's move under the SpaceX umbrella

* Meta's Muse Spark 1.1 and its first paid frontier API

* The price war running across all three launches

* SK Hynix's record IPO and the money flowing into compute

Sources

* OpenAI launches its new family of models with GPT-5.6

* SpaceX's xAI releases Grok 4.5

* Meta opens its Model API with Muse Spark 1.1

* SK Hynix raises $26.5B in the biggest foreign IPO in US history

* SambaNova draws $1B at $11B valuation

* Reflection inks $1B compute deal with Nebius

Episode 49 unpacks a week where three companies answered the same question in three different ways: what does AI actually cost, and who pays for it. Anthropic shipped Claude Sonnet 5 at $2 per million input tokens, close to flagship Opus 4.8 quality, betting most work doesn't need the expensive model to get expensive-model results. Meta started renting out its spare AI compute as a new business called Meta Compute, chasing the same cloud dollars as AWS and Google. Tesla went the other way, capping employee AI tool spending at $200 a week months after ranking engineers by how many tokens they burned.

What's Covered

* Anthropic's Claude Sonnet 5 launch and its near-Opus pricing bet

* Meta Compute: turning spare AI infrastructure into a product

* Tesla's $200-a-week cap on employee AI tool spending, and the Grok exception

Sources

* MacRumors: Anthropic releases Claude Sonnet 5

* TechCrunch: Meta, like SpaceX, looks to turn excess AI compute into cash

* Electrek: Tesla caps employee AI spending at $200/week

This week's episode digs into the moment the AI race moved from pure software to the hardware and control beneath it. OpenAI and Broadcom reveal Jalapeño, OpenAI's first custom chip. Anthropic takes a 29-million-query theft of Claude to the White House. And GPT-5.6 arrives gated to about twenty vetted organizations. Along the way: the tool that turns messy PDFs into clean data, and a prompt move that lets a cheap model punch above its weight.

What's Covered

* OpenAI and Broadcom's Jalapeño chip and the move off Nvidia

* Anthropic's distillation accusation against Alibaba

* GPT-5.6 Sol, Terra, and Luna, and the government-gated rollout

* LlamaParse, the tool of the week

* Metaprompting, the prompt of the week

Sources

* OpenAI unveils its first custom chip, built by Broadcom (TechCrunch)

* Anthropic claims Alibaba illicitly distilled Claude (Tom's Hardware)

* Previewing GPT-5.6 Sol, Terra, and Luna (OpenAI)

* LlamaParse (LlamaIndex)

* Anthropic Economic Index, June 2026

* Users shift from tokenmaxxing to efficiency (CNBC)

* US clears Mythos 5 for 100+ organizations (TechCrunch)

* Nobel laureate John Jumper leaves DeepMind for Anthropic (TechCrunch)

* Google delays Gemini 3.5 Pro to July (Crypto Briefing)