AI

LWiAI Podcast #252: GPT 5.6, Grok 4.5, Nemotron-Labs-Diffusion, and AI 2040

OpenAI rolls out GPT 5.6 with Sol and Luna variants, SpaceX AI launches Grok 4.5 as a low-cost coding model, and AI 2040 proposes US-China coordination on AI safety.

LWiAI Podcast #252 covers OpenAI's GPT 5.6 rollout, SpaceX AI's low-cost Grok 4.5, Meta's Muse Spark 1.1, NVIDIA's Nemotron-Labs-Diffusion, and AI 2040's proposal for US-China coordination. The episode highlights intensifying model competition and policy discussions.

The 252nd episode of Last Week in AI covers a packed week of AI news, from major model releases to policy proposals that could shape the future of artificial intelligence.[reference:84]

OpenAI's GPT 5.6 Family

OpenAI publicly rolled out GPT-5.6, including Sol and Luna variants, and rebranded its desktop agentic coding product as ChatGPT Work.[reference:85] The release came amid disputed claims about whether the US government effectively green-lit and delayed the release, and concerns about inconsistent, ad hoc frontier-model oversight and jailbreakability.[reference:86]

SpaceX AI's Grok 4.5

SpaceX AI launched Grok 4.5 as a very low-cost, Opus-class coding model with minimal safety documentation.[reference:87] The model is priced aggressively, with rates of 2/2/6 per million tokens, signaling a shift toward specialized agentic models.[reference:88]

Meta's Muse Spark 1.1

Meta released Muse Spark 1.1 with aggressive pricing, large coding and cyber benchmark gains, and a lengthy safety evaluation.[reference:89] Meta also previewed Muse Video and rolled out Muse Image before quickly backtracking after backlash over easy generation of images of public Instagram accounts.[reference:90]

Nemotron-Labs-Diffusion

NVIDIA released Nemotron-Labs-Diffusion, adding to the growing ecosystem of diffusion models.[reference:91]

AI 2040: A Policy Roadmap

AI 2040 proposes US-China coordination to slow progress until alignment improves.[reference:92] The proposal reflects growing recognition that international cooperation may be necessary to manage the risks of advanced AI.

Chinese open-source models grew to over 30% of weekly OpenRouter tokens as cost pressure increased, alongside discussion of risks like potential insider threats.[reference:93] The model competition is increasingly shifting from general capabilities to specialized agentic scenarios.[reference:94]

The Bigger Picture

The episode captures a moment of intense competition and growing policy concern. With multiple frontier models launching in quick succession, the AI industry shows no signs of slowing down. But with each new release comes new questions about safety, oversight, and the long-term trajectory of AI development.