GoKawiil
Tech News
Anthropic has launched Claude Fable 5.2 and Mythos 5.2, successors to the Fable 5.1 update released in September. The company says the new models improve on coding, reasoning, and vision tasks, run up to 30% faster than their predecessors, and cost $8 per million input tokens and $40 per million output tokens.
Google has launched Gemini 4 Carbon, which Artificial Analysis now ranks as the top AI model on its Intelligence Index V5 suite, narrowly surpassing Anthropic's Claude Opus 5.5 by one point and besting OpenAI's GPT-6 Astra and Meta's Muse Wonder Full. The model also leads a separate Artificial Analysis test measuring hallucination rates and factual recall, though its score of 54 out of a possible 100 shows even the top model remains far from fully reliable. Google is initially rolling the model out to cyber defenders via its Fairwind program, with broader access to follow later.
Developer Brandon Thomas built Artcraft, a set of free, open-source clones of Adobe Creative Cloud apps including a Photoshop substitute called Photocraft, using Claude Opus 5.5 to reverse-engineer the software in Rust. Thomas says the move was prompted by frustration with Adobe's subscription cancellation fees, and the apps mimic Adobe's interfaces closely enough to require little relearning. Testing found Photocraft handles basic tasks like resizing and masking well but suffers from bugs, instability, and an inability to open native PSD files.
Artificial Analysis reports that Google's new Gemini 4 Carbon model has topped its Intelligence Index v5 with a score of 59, ranking first in 4 of 9 component evaluations including HLE-Diamond and AA-Omniscience. The model also placed second on the firm's Coding Agent Index, trailing only Anthropic's Claude Opus 5.5, and showed sizable gains in agentic knowledge-work benchmarks over its predecessor, Gemini 4 Argon. The model is currently in limited rollout with a 50% launch discount and not yet publicly available.
Moonshot AI released Kimi K3.1 Pro, an open-weight 3T-class model built on Kimi K3 with expanded post-training compute, offering a 1-million-token context window and gains in coding, 3D creation, financial research and agentic tasks. A faster variant, K3.1 Pro Highspeed, runs 2-3 times quicker and both are live now on Kimi.ai, Kimi Work, Kimi Code and the Kimi API, with full model weights due by November 4.
Anthropic introduced Claude Duet, an AI agent app built on its Claude Opus 5.5 model targeting professionals and small businesses rather than personal errands. The app can generate daily briefings, build dashboards, suggest actions from user data, and complete background tasks via its own virtual machine before seeking approval. It connects to over 6,000 plugins in the Claude Marketplace and can make purchases through Stripe Link or Shopify Shop Pay, though it won't trade securities or transfer money between accounts.
Anthropic introduced Claude Duet, a new AI agent built on Claude Opus 5.5, launching as a research preview on iOS, Android, and the web. The agent monitors connected apps, sends daily briefings, flags urgent items, and can handle tasks ranging from scheduling to building dashboards and prototypes, pausing for user approval before sensitive actions like payments. Access currently requires joining a waitlist and holding a Pro or Max Claude subscription in the US.
During the current Steam Next Fest, players have noticed a sharp rise in demos featuring AI-generated art or code, prompting complaints on Reddit. One user, ScalesGhost, asked Valve to let players filter out games disclosed as containing AI-generated content and exclude specific tags like 'Incremental Games' from the Discovery Queue. Other users, including EmmaBooney and aesoprowwy, said the flood of 'AI slop' has made it harder to find human-made games and hope for lower engagement on such titles.
A Show HN post introduces Playground, a site that collects browser games built with AI tools like GPT and Claude models. Creators submit games via a form including a play link, cover image, device compatibility and category tags, and each submission is manually reviewed before being listed. The service is free to submit to.
A fan named supermochi64 used the AI model Claude Opus 5.5 alongside GPU-based font testing to decode the long-illegible 32x32 pixel plaque text in Super Mario 64's castle courtyard. The method, which tested millions of font and word combinations against the texture, scored 'Eternal star / We hope so' as a far better match than the popular fan theory 'L is real 2401.' The analysis also identified the likely font as Palatino Bold Italic from Adobe's 1990 Mac release.
A new benchmark called StarSkirmish Bench v0.2 tests flagship LLMs from OpenAI, Anthropic, Google and Meta by giving each one hour to write a Protoss bot in C++ for StarCraft: Brood War, then scoring win rates against human-written and demo bots. GPT-6 Astra topped the leaderboard, with GPT-6.1 Sol second, while Gemini 4 Argon, Claude Opus 5.5 and Claude Fable 5.1 scored closely behind between 37 and 39. Even the top model, GPT-6 Astra, won only 7% of games against the strongest human-written bots like Stardust, PurpleWave and BananaBrain.
Ultramarine 1 Flash is a new model priced at $0.13 per task on internal benchmark PI Knowledge Work, which the company says correlates strongly with the Artificial Analysis Intelligence Index. It is built on GLM-5.3 Flash and generates 6.5% fewer tokens than that base model while outscoring Claude Fable 5 (max) and GPT-5.6 Sol (max), which cost $8.19 and $2.09 per task respectively. Pricing on the Ultramarine API is $0.15 per million input tokens, $0.03 for cached input, and $0.50 for output.