AI Daily · Oct 8, 2026: OpenAI rolls out GPT-6 and Intelligent UI to all ChatGPT users as Anthropic launches Claude Haiku 5.5 at 75% lower cost
OpenAI announced the GPT-6 and Intelligent UI rollout to all ChatGPT users, plus a one-time usage reset for paid accounts. Anthropic launched Claude Haiku 5.5 at roughly 75% lower running cost than Haiku 4.5, and halved Claude Sonnet 5.5 cache-read pricing to $0.10 per million tokens. Step 5 Preview
Pick a date
-
OpenAI rolls out GPT-6 with Intelligent UI across ChatGPT subscription tiers — GPT-6 with Intelligent UI begins rolling
OpenAI announced that GPT-6 and an Intelligent UI layer are rolling out to everyone in ChatGPT, delivering faster interactive answers and on-the-spot task tools.
-
Anthropic launches Claude Haiku 5.5, 75 percent cheaper than Haiku 4.5 — Anthropic released Claude Haiku 5.5, its cheape
Anthropic introduced Claude Haiku 5.5 with average running costs about 75% below Haiku 4.5, now available on major cloud platforms and through Cursor, LiteLLM, Requesty, Command Code, B.AI, OpenCode and other channels.
-
Claude Sonnet 5.5 cache read price cut to $0.10 per million tokens — Cache reads for Claude Sonnet 5.5 dropped from $0.2
Cache reads for Claude Sonnet 5.5 dropped from $0.20 to $0.10 per million tokens, roughly 20% cheaper for long-running workloads; the post also lists Haiku 5.5 tiered pricing.
-
OpenAI announces 40M active Codex and ChatGPT Work users, banked reset for paid accounts — Tibo of OpenAI reported a new
OpenAI loaded a banked usage reset into all paid accounts, while noting the approve-for-me auto-review feature consumes between 2-10% of plan usage.
-
Step 5 Preview launches on OpenRouter with free week access — Step 5 Preview, StepFun's sparse MoE flagship (27B active
StepFun's flagship Step 5 Preview launched on OpenRouter at $1.00/M input and $2.70/M output with 50% cache-hit discount, plus a one-week free rollout across OpenRouter, opencode, cline, Nous Portal and other partners.
-
Google opens SynthID Detector verification portal globally with partner support — Google's SynthID Detector is now publi
Google's SynthID Detector is now publicly available in English worldwide, letting anyone check whether online media contains SynthID watermarks, including content from multiple partner tools.
-
Google Releases EmbeddingGemma 2 Multimodal Embedding Models — Google released EmbeddingGemma 2, its first native multim
Google released EmbeddingGemma 2, its first native multimodal embedding model built on Gemma 4 under Apache 2.0, embedding text, code, images, audio, and video across 100+ languages into a single vector across four parameter sizes.
-
Perplexity open-sources pplx-embed-v2-late retrievers under MIT — Perplexity released pplx-embed-v2-late on Hugging Face
Perplexity released pplx-embed-v2-late on Hugging Face under the MIT license in 0.6B and 9B sizes, supporting cross-modal retrieval of text, images and pages.
-
OpenRouter Adds GPT-6 Luna Decisions With Decisions API Pricing — OpenRouter launched GPT-6 Luna Decisions, powered by O
OpenRouter launched GPT-6 Luna Decisions, powered by OpenAI's Decisions API in public beta, priced at $0.10 per million input tokens with free output, a 1M context window, and typed answers with probabilities.
-
Glyph Cluster stealth reasoning model free on Command Code plans — Command Code announced its new stealth model Glyph Cl
Command Code announced its new stealth model Glyph Cluster, a reasoning model with a 256K context window, available for free on all plans via a specified command.