Models briefing

Open-sourcing AstaBrief, the fast report-generation model in Asta

Models

InstructMesh lets users refine AI-generated 3D models before fabrication

The tool combines a 3D generator and a language model to help users identify and repair design flaws.

Language Discrimination Improves Linguistic Learning in Multilingual Speech Models

ChatGPT can now virtually try on clothes for you

Models

Upstage releases Solar Mini 4, a proprietary model with 3B active parameters

Artificial Analysis reports a score of 24 and lower per-token prices than Solar Pro 3, alongside higher measured costs per task than GPT-6 Luna (max).

Introducing Clef: our open-source decision models, and new RL fine-tuning platform

Models

Study tests what pretraining and midtraining contribute to reward learning

Experiments and analysis examine how source prediction and reward adaptation interact in models learning sequential tasks and memory.

Models

Open TTS Leaderboard separates streaming speed from voice-quality measures

The evaluation pairs latency and intelligibility metrics with speaker similarity and sample listening, without treating any single measure as a verdict on voice quality.

Models

Raycast 2.6 adds preferences for automatic model selection

Raycast’s latest release offers three Auto model preferences with different trade-offs in speed, credit use and model capability.

Models

GoldiMask changes how diffusion language models use training context

A new fine-tuning method selects visible context tokens and weights prediction targets, with reported gains across several model and dataset settings.

The Den frees up 10-15 hours a week to grow with ChatGPT Work

Gemini 4 Argon: our next era of frontier intelligence

Identify AI model overuse with User Insights

Disrupting a coordinated model-distillation campaign

Amazon Bedrock expands Claude model availability to in-country inferencing in India

Introducing Anthropic models on Amazon Bedrock for in-region inference in Seoul and Singapore

Bring near-Astra intelligence to everyday work with GPT-6.1 Sol on Amazon Bedrock

AI Native by Design: Lessons Learned from Building NVIDIA TensorRT Model Connect

Models

GPT-6.1 Sol begins rolling out across GitHub Copilot

The model is available to four Copilot plans, with access expanding gradually across supported coding tools.

Models

Claude Sonnet 5.5 rolls out across GitHub Copilot

The model is available to users on supported Copilot plans, with access rolling out gradually across clients.

Models

Anthropic releases Claude Sonnet 5.5, a faster and cheaper model that nearly matches Opus 5.5 on several benchmarks

Sonnet 5.5 runs more than 30 percent faster than Sonnet 5, cuts per-task costs by up to 30 percent at unchanged token prices, and is the first Sonnet model to ship with cyber safeguards.

Models

Raycast 2.5 lets extensions add models to AI Chat

Users can install compatible extensions, complete setup and select the extensions’ models in AI Chat.

Models

Langfuse adds TypeSafe’s Jev as an evaluator

The decision model scores observations using typed questions, with options for categorical and numeric results.

Models

OpenAI and OATS host ChatGPT workshops

OpenAI Academy and OATS are collaborating to provide free, hands-on ChatGPT workshops for older adults across 10 U.S. cities.

Models

OpenAI publishes misalignment reporting framework and six incident reports

OpenAI released a framework for tracking, investigating and disclosing model misalignment, along with six reports on unexpected or concerning behavior it said it observed during training and evaluation over the past six months. The company said no industry-wide disclosure standard currently exists.

Models

TypeSafe AI exits stealth with $40M seed and Jev, a model that outputs decisions

San Francisco startup TypeSafe AI emerged from stealth on Sept. 15 with $40 million in seed funding led by DCVC and a first model, Jev, that returns typed decisions with probabilities for software to consume rather than generating natural-language text. Access is through an early-access waitlist.