YFarmX logoYFarmX

AI Hub

Explore the future
of artificial intelligence

Models, agents, infrastructure and policy: what actually changed, and what it means.

64
AI Index records
174
AI explainers
117
AI stories published
33
verified events ahead
NVDANVIDIA$237.07up3.81%
MSFTMicrosoft$517.32up0.86%
GOOGLAlphabet$344.84up0.22%
AMDAMD$641.19up4.81%

Nasdaq · 02 Oct 2026, 14:19 UTC · refreshed every six hours

Model of the weekGoogle’s Gemini 4 Argon supports million-token outputs and starts with cyber defenders

Latest AI news

All AI news →
Editorial collage headed Tencent, with a long row of server racks receding through a data hall, the nearest lit blue and tagged 100,000, and the Tencent and Oracle logos; the subtitle reads reported lease, 100,000 AI chips.

Lead story

Tencent rents 100,000 AI chips from Oracle, the FT reports

The Financial Times reported on 30 September 2026, in a report Reuters carried but could not verify, that Tencent has agreed a five-year lease for about 100,000 advanced AI chips in several Oracle data centres in south-east Asia, estimated at about $7bn with about 30% paid upfront.

7 min read◉ audio

110 more AI stories in the archive →

State of play

TIME-SENSITIVEupdated 2 October 2026

LLMsGoogle announced Gemini 4 Argon on 30 September with a one-million-token output limit, starting access with trusted cyber defenders and testers. Its announced introductory API prices are $2 input and $10 output per million tokens. Anthropic ships Claude Opus 5.5 on 22 September at $4 and $20 per million tokens, and OpenAI answers the same day with GPT-6 Sol and GPT-6 Luna at half the GPT-5.6 prices. On Artificial Analysis's Intelligence Index v4.3, read on 2 October, Opus 5.5 leads on 58, Sonnet 5.5 has sat second on 56 since 28 September, Fable 5.1, GPT-6 Astra and Gemini 4 Argon are level on 53, and GPT-6.1 Sol scores 52. The frontier text picture sits in the dated log below.Image GenerationOpenAI's Images 2.5 pair still holds the top two arena seats: Sunburst and Flare lead both the text-to-image and image-edit boards with GPT Image 2 third on each, so OpenAI holds the top three seats on both; Microsoft's MAI-Image-2.6 is fourth on text-to-image (arena.ai boards read 2 October; text-to-image dated 24 September, image edit dated 29 September). Reve 2.1 still shows fifth on that text-to-image board, but Reve stopped generating images on 27 September and gives users until 31 October to download their work. Images 2.5 adds sketch-guided creation, image comments and more consistent editing, with Flare built for speed and Sunburst for precision.Video GenerationGoogle's Gemini Omni 1.1 Flash and Omni Flash lead text-to-video on 1516 and 1513 Elo, FLUX 3 Video is third on 1493 with xAI's Grok Imagine Video 1.5 agent mode a point behind, and Wan 3.0 has slipped to sixth (arena.ai board dated 21 September, read 27 September); MiniMax H3 holds image-to-video and Wan 3.0 video editing. OpenAI's Sora 2 API shut down on 24 September, closing the line.
Open AI State of Play: all 7 areas

Explore the AI desk

Listen: AI briefings

All briefings →