Qoder Launches 'Fast' Mode With 3s Responses at Lower Credit Cost
Qoder introduces a new 'Fast' performance mode delivering 3-second responses with reduced credit consumption while maintaining the same output quality.
Global AI Digest
-
-
MiniMax M3 Runs on NVIDIA Vera Rubin NVL72 via vLLM, Claiming 7.8x Throughput Over GB200
MiniMax M3 is now running on NVIDIA Vera Rubin NVL72 through vLLM, with early results showing over 7.8x the throughput of GB200 on AgentX benchmark.
-
Opencode Rewrote OpenTunnel Backend in Rust and Migrated Off Cloudflare to Fly.io
Opencode rewrote the OpenTunnel backend in Rust and moved it from Cloudflare to 10 Fly.io regions, with a graceful migration that kept users' tunnels connected.
-
ChatGPT Subscriptions Now Work as Devin Subscriptions
Personal ChatGPT plans (Free, Plus, Pro) can now be linked to Devin, with all GPT model usage drawing down from the user's ChatGPT plan quota.
-
Dots Can Now Delegate and Follow Up on Codex Work
Dots received a major upgrade allowing them to start work in Codex, follow up on existing threads using ChatGPT conversations and automations, and review schedu…
-
ChatGPT Weekly Feature Drop: GPT-6, Free Finance Access, and More
A roundup of ChatGPT's Oct 9 feature drop, including GPT-6 and Intelligent UI availability for all users and finance tools opened to free users in the US.
-
Opencode Server and TUI Rewritten via a Single Prompt Orchestrating Subagents for $500
Opencode 团队用从 iOS 应用发出的单条提示词，由 Claude Opus 5.5 编排子代理，重写了 opencode2 的服务端与 TUI，总成本约 500 美元。
-
Nano Banana 2.1 Now Available to All GMI Cloud Users With Improved Editing Features
GMI Cloud announces Nano Banana 2.1 is available to all users, featuring better visual design, mask-based editing, and more natural-looking images, accessible v…
-
Microsoft-Decision-1 Launches on OpenRouter with 32K Context at $0.042/M Input
Microsoft-Decision-1 is now available on OpenRouter, claiming top accuracy on 36 blind benchmarks, 4.5x faster than the runner-up, priced at $0.042/M input with…
-
Anthropic Begins Publishing Frequent Model Behavior Reports, Details Four Unintended Claude Behaviors
Anthropic announced it will publish model behavior reports more frequently beyond system cards, with the first describing four behaviors where Claude acted on r…
