Thanks for the support @AndrewYNg! Completely agree, faster token generation will become increasingly important as a greater proportion of output tokens are consumed by models, such as in multi-step agentic workflows, rather than being read by people.
Can’t get to Miami for the GTA VI launch? Two AI models built the city for you
We asked Claude Fable 5.2 from @AnthropicAI and GPT-6.2 Astra from @OpenAI to build Miami as it looks this week, dressed as Vice City for the GTA VI launch
➤ One prompt each, in the AA-3D sandbox with no internet access
➤ Both lined Ocean Drive with real hotels the prompt never named, including the Colony and the Avalon
➤ Each scene can be explored by car, speedboat or blimp, at sunset or at night
GPT-6.2 Astra ranks #1 on AA-3D, our benchmark measuring how well models can build interactive 3D experiences. Humans preferred these scenes in 139 of 152 head-to-head votes against other models
Microsoft-Decision-1.1 scores 68 on the Artificial Analysis Decision Index v2.1
@Microsoft has released Microsoft-Decision-1.1, an update to their existing decision model, Microsoft-Decision-1. It remains priced at $0.042 per one million input tokens.
Key results:
➤ Microsoft-Decision-1.1 is the #8 decision model. On the Artificial Analysis Decision Index, Microsoft-Decision-1.1 scores 2 points ahead of Elephant 9B and Jev 1.14 but 1 point behind d1.2.
➤ Microsoft-Decision-1.1 improves by 6 points over its predecessor. Microsoft-Decision-1.1’s score increases by 6 points to 68 from Microsoft-Decision-1’s 62, improving across all 6 individual evaluations that make up the Artificial Analysis Decision Index.
➤ Microsoft-Decision-1.1 is the #5 fastest model on a per task basis. Microsoft-Decision-1.1 has a Time per Task on the Artificial Analysis Decision Index of 123 ms, faster than d1.2 (150 ms) but not as a fast as Mammoth 27B (106 ms). It is more than 2x as fast as Microsoft-Decision-1 (274 ms).
GPT-6.2 Astra is the new leader in AA-Robotics v0.1, scoring 79, up 10 points on GPT-6 Astra and ahead of GPT-6.2 Sol (78), Claude Fable 5.2 (75) and Gemini 4 Hydrogen (58)
Controlling a real robot arm through Codex, it passed six of our eight physical tasks, partly completed the pouring liquid, and failed only to measure the box.
MAI-Image-2.7 ranks #1 for 4 of 10 Image Editing use cases: Architecture & Real Estate, Consumer, Frontier, and Marketing & Advertising
Use cases measure the workloads end users run. We measure the same 10 use cases for Image Editing and Text to Image.
In Text to Image, MAI-Image-2.7 ranks #2 for Marketing & Advertising and for Productivity & Knowledge Work. MAI-Image-2.7-Flash's highest use-case ranks are #3 for Frontier and Animation & Gaming edits, and #5 for Productivity & Knowledge Work in Text to Image.
➤ Marketing & Advertising covers ad creatives, campaign visuals, posters, and brand imagery.
➤ Architecture & Real Estate covers interior and exterior visualization, floor plans, and staging.
➤ Frontier covers capabilities at and beyond the current frontier.
MAI-Image-2.7 ranks #1 on 4 of 7 editing actions and on Knowledge in Text to Image. MAI-Image-2.7-Flash's strongest editing actions are Scene & Style Edit (#3) and Enhancement & Restoration (#4)
Our Image Arena taxonomy measures 7 editing actions for Image Editing and 9 capabilities for Text to Image. Editing actions measure what an edit asks the model to do. Capabilities measure what a prompt asks the model to get right.
MAI-Image-2.7's largest gaps to the leader are Object-Level Edit in Image Editing and Lighting in Text to Image.
➤ Text or Symbol Edits covers editing text, typography, logos, and symbols within the image.
➤ Enhancement & Restoration covers denoising, retouching, artifact removal, and repair of damaged or degraded images.
➤ Knowledge covers real landmarks, species, and domain facts across science and common sense.
MAI-Image-2.7 and MAI-Image-2.7-Flash both set new points on the Quality vs Price frontier, in Image Editing and in Text to Image
Released November 17, MAI-Image-2.7 and MAI-Image-2.7-Flash are Microsoft AI's new models for text to image generation and image editing, now in public preview. MAI-Image-2.7 follows MAI-Image-2.6 and the editing-only MAI-Image-2.7-Preview from October 27. We benchmarked both models through Microsoft's API with square outputs.
➤ Ranks: MAI-Image-2.7 is #1 on AA-Image-Editing v2.1 and #4 on AA-Image-T2I v2.1. MAI-Image-2.7-Flash is #7 and #11
➤ Frontier: MAI-Image-2.7 is the new top of the Image Editing frontier, replacing GPT Image 2.6 Sunburst (max). MAI-Image-2.7-Flash sets a new lower-price point on both boards, pushing Seedream 5.5, MAI-Image-2.6-Flash and Canva Craft off the frontier
➤ Price: $34.82 per 1k images for MAI-Image-2.7, down from $38.91 for MAI-Image-2.6. $10.24 per 1k images for MAI-Image-2.7-Flash, down from $19.46 for MAI-Image-2.6-Flash. Both at Microsoft Foundry public preview rates
➤ Availability: Microsoft Foundry, MAI Labs (including the new Canvas experience, powered by MAI-Image-2.7-Flash), and providers fal, Genspark, OpenRouter and Vercel
Microsoft AI has launched MAI-Image-2.7, the new #1 in AA-Image-Editing v2.1 at one sixth the price of the four models behind it
MAI-Image-2.7 and MAI-Image-2.7-Flash both take new places on the Quality vs Price frontier in Image Editing and Text to Image, at $34.82 and $10.24 per 1k images.
Both perform best in Image Editing: MAI-Image-2.7 ranks #1 for Text or Symbol, Scene & Style and Enhancement & Restoration edits, and MAI-Image-2.7-Flash ranks #3 for Scene & Style edits.
Congratulations to @MicrosoftAI on the launch of MAI-Image-2.7 and MAI-Image-2.7-Flash!
See below for our detailed analysis of both models in Image Editing and Text to Image 🧵
OpenAI leads Harvey LAB-AA v1.1 with Astra for Law 6.2, a specialized GPT-6.2 Astra configuration with an integrated legal search index
OpenAI describes Astra for Law 6.2 as a model configuration of GPT-6.2 Astra that pairs the model with an expanded legal search index and instructions for researching a question, weighing the authorities, and writing the answer.
Our implementation of Harvey’s LAB benchmark tests real-world legal work across 120 private tasks spanning 24 practice areas, penalizing hallucination in its headline metric, Hallucination-Gated All-Pass Rate. Astra for Law leads Harvey LAB-AA v1.1 with a score of 20.8%. GPT-6.2 Astra (max) comes in second with 15.0%.