Claude Fable 5 Outperforms GPT-5.5 on FrontierMath Benchmark
Anthropic's Claude Fable 5 achieved 88% accuracy on FrontierMath's hardest tier 4 problems, outperforming GPT-5.5 by 13 points in June 2026.
Model Release
214 articles
New AI models as they ship — large language models, multimodal systems, and open-weight releases. Capabilities, context windows, pricing, and availability, summarized so you can compare what is actually new.
LATEST NEWS
-
-
Moonshot's Kimi K2.7 Code Offers 12x Lower Token Pricing Than GPT-5.5 and Claude
Moonshot AI's Kimi K2.7 Code is priced at $0.95 per million input tokens and $4.00 per million output tokens, undercutting GPT-5.5 and Claude by up to 12x.
-
Anthropic's Claude Fable 5 Costs Twice as Much for 5.7 Percent More Performance
Anthropic's Claude Fable 5 scores 64.9 in the Artificial Analysis Intelligence Index, outperforming GPT-5.5 by five points, but costs double the price of its predecessor.
-
OpenAI's Super App Strategy Led by Tibo Sottiaux
OpenAI is developing a unified AI platform, calling it a 'super app,' with Tibo Sottiaux leading the effort. The project aims to merge Codex and ChatGPT into a single tool, with plans to release updates in coming weeks.
-
Apple's Siri AI Overhaul and SpaceX IPO Highlighted in Uncanny Valley Podcast
Apple's rebranded Siri AI and SpaceX's record-breaking IPO are central to this week's podcast, with Meta removing face recognition code after a WIRED report.
-
Avataar AI Launches Varya, a Culturally Aware Video AI Model
Avataar AI’s Varya video model generates 5-second clips 10 times faster than its predecessor, with pricing 20x lower than global alternatives.
-
Apple Introduces AI-Driven Photo Editing in iOS 27
Apple's iPhone will get AI-powered photo editing tools in iOS 27, which will arrive later this year. The features include Extend and Spatial Reframe, which generate fake pixels to alter images.
-
Anthropic Admits Mistake in Claude Fable 5 Throttling Rival AI Researchers
Anthropic reversed its plan to secretly throttle AI researchers using Claude Fable 5 after backlash, citing a wrong tradeoff in its policy.
-
Google Releases DiffusionGemma, Open Model That Generates Text From Noise
Google has released DiffusionGemma, an open-source language model that generates text through diffusion, achieving speeds up to four times faster than traditional models on dedicated GPUs.
-
Google releases DiffusionGemma, a new open AI model with 4x speed boost
Google DeepMind has released DiffusionGemma, a new open AI model that generates text 4x faster than similar autoregressive models, according to a June 2026 report.
-
DeepMind Releases DiffusionGemma, 4x Faster Text Generation Model
DeepMind announced DiffusionGemma, an experimental open model that generates text up to 4x faster on GPUs, with performance benchmarks reaching 1000+ tokens per second on an NVIDIA H100.
-
Anthropic Releases Claude Fable 5, the First Mythos Model
Anthropic has released Claude Fable 5, the first publicly available Mythos model, which tops nearly every benchmark but faces criticism over cost and safety filters.