November 2026.
10 posts from the blog, newest first.
Archive
· 5 min read
Claude Fable 5.2: price, benchmarks and how to try it
Anthropic’s Claude Fable 5.2 and Mythos 5.2 cost $8/$40 per million tokens, with a 1M context. What’s new, benchmarks, fast mode and how to try it.
Claude Fable 5.1Claude Opus 5.5+1
· 4 min read
K-EXAONE 2.5: free API, 750B specs and benchmarks
LG AI Research opened a free API for K-EXAONE 2.5, a 750B model (31B active) with 512K context. Specs, LG-reported benchmarks and how to try it.
DeepSeek-V4.1-FlashKimi K3.1 Pro+1
· 5 min read
Inkling 2 Small explained: specs, price, open weights
Thinking Machines’ Inkling 2 Small is a 594B MoE with 17B active, image input and $0.30/$1.00 Tinker pricing. Specs, benchmarks and the open-weights plan.
Inkling 2 SmallKimi K3+1
· 5 min read
Gemini Omni 1.2 Flash: 15-second video, price, how to try
Google’s Gemini Omni 1.2 Flash is GA: 15-second clips, edits by follow-up prompt, $0.07/s at 720p, how to try it and what people are building.
Gemini Omni 1.2 Flash
· 5 min read
GPT-6 Cyber explained: Daybreak Red, price, scores
OpenAI launched GPT-6 Cyber on 5 November 2026: Daybreak Red access, API pricing, ExploitGym scores, Firefox and Linux findings, and how to apply.
GPT-6 CyberGPT-6 Astra
· 5 min read
GPT-Image-2.6 Flare and Sunburst explained
OpenAI shipped GPT-Image-2.6 Flare and Sunburst on 4 November 2026: faster images, GIFs, SVG export, markup edits, pricing and how to try them.
GPT-Image-2.6 FlareGPT-Image-2.6 Sunburst
· 5 min read
P-H3-Image open weights: Pruna’s MiniMax-H3 image model
Pruna released P-H3-Image on 3 November 2026: an open-weight text-to-image and editing model built on MiniMax-H3, with Diffusers, ComfyUI and license details.
· 4 min read
Hydrogen-1.0 explained: Deep Principle’s science model
Deep Principle’s Hydrogen-1.0 is an open-weight science model post-trained from GLM-5.3. What’s new, self-reported scores, how to try it, and who it’s for.
Hydrogen-1.0GPT-6 Astra+1
· 5 min read
LongCat-2.5 explained: Meituan’s 1.6T agentic model
LongCat-2.5 is Meituan’s 1.6T-parameter multimodal model for agentic coding. What’s new, benchmarks, API pricing, how to try it and early builds.
LongCat-2.5Claude Opus 5+1
· 5 min read
Apple Intelligence in China: Qwen, features and dates
Apple Intelligence comes to China with Alibaba’s Qwen in Siri and Writing Tools. Beta from Nov 4, iOS 27.3 in December. Features, iPhones, what’s missing.