Ashlr Brief.

Andy Jassy told analysts that frontier model competition will produce several comparably capable systems. Bedrock already routes to more than 100 models through one API, and most traffic now runs on Trainium.

Mistral AI has made the weights of its 1-trillion-parameter multimodal model available under open release, enabling sovereign control and local deployment options.

Subscribers to $100, $200, or $500 ChatGPT plans now receive equivalent monthly credits usable with the OpenAI API, following a similar benefit from Anthropic.

Nvidia launched an open-weight 30B mixture-of-experts transcription model that accepts plain-language prompts for context, speaker labels, translation, and audio questions. It supports 40 languages and runs on a single GPU via vLLM or NeMo-Speech.cpp.

GitHub released Copilot Fusion, a generally available system that selects and orchestrates models for coding tasks. Enterprise admins can restrict the model pool while developers choose effort levels instead of individual models.

Anthropic's Claude Code v2.1.311 introduces CLAUDE_CODE_EVEN_MORE_DANGEROUSLY_SKIP_PERMISSIONS, removing the final guardrails on commands like rm. The change follows the earlier --dangerously-skip-permissions flag and bypassPermissions mode.

Genie 4 renders real-time simulated environments with physics and audio. Gemini models steer the scenes so agents can rehearse multi-step tasks before deployment. Access remains limited to trusted testers and planned Google AI Ultra subscribers.

OpenAI launched GPT-6.1 Terra, a mid-tier model for computer use and visual reasoning tasks. It costs roughly half as much as GPT-5.6 Terra while running faster on agent workflows.

OpenAI announced AI-produced results on local smoothing, the Exponential Time Hypothesis and quantum field theory. Experts are now checking them.

Anthropic filed its S-1 for an IPO under the ticker ANT. Share count and price are not set yet. Here is what the filing does and does not tell you.

OpenAI released GPT-6.1 Terra, a fast model for computer use and everyday work at $1 per million input tokens and $5 per million output tokens.

AWS made its Continuum agent generally available. The service finds, validates, and proposes fixes for code vulnerabilities in repositories while checking the AWS environment.

Bernstein Research estimates only about 35 percent of announced US data center capacity will be built. Multiple sources point to electricity constraints and speculative projects inflating reported demand.

Jack Dorsey’s Block introduced Dessa, a speech-to-speech model trained only on human conversation. It runs locally on high-end Macs and is entering limited safety researcher access.

AWS released general availability of training sessions in SageMaker AI serverless model customization. Builders can now write their own post-training loops, including custom loss functions and multi-turn RL, while SageMaker handles GPU allocation, distributed execution, and fault recovery.

Claude Duet is a research preview for US Pro and Max subscribers through a waitlist. Anthropic says it briefs users and works across connected apps.

IBM’s ALTK-Evolve turns past agent runs into a playbook. IBM reports more consistent support-task results, with higher input-token costs.

Mistral’s Voxtral 3 combines speech, reasoning and tools. API hosting and self-deployment need different checks before a live pilot.

Google and Hugging Face announced native TPU support for PyTorch. Compatibility, warm-up time and cost still need workload-specific checks.

OpenAI published coding evaluations for less common languages. Its migration tests ask whether old software’s behaviour survives a rewrite.

Anthropic proposed an equity-backed public-benefit fund. Stockholder approval, governance and future distributions remain separate questions.

MiniMax says M3.1 Flash can run agent workloads for less. Its posted token prices are low; the useful comparison is what you pay for a completed job, including retries.