Requesty Adds Per-Key Agent Visibility and One-Click Key Revocation
Requesty introduced a feature that shows which agents (Claude Code, Cursor, Codex, curl) are using each API key, helps detect leaked keys, and allows revoking t…
Global AI Digest
-
-
Qoder Launches $10/Month Lite Plan With 1,000 Credits
Qoder introduced a Lite subscription at $10/month with 1,000 credits, Flash models at 0.1x cost for everyday tasks, and access to Qwen3.8-Max, DeepSeek-V4-Pro a…
-
DeepSeek Open-Sources DeepInfer Inference Framework and Serving Recipes
DeepSeek has open-sourced DeepInfer, its in-house inference framework, along with deepseek-recipe and other components for serving DeepSeek models, aiming to he…
-
Qwen3.8 Flash-Next Hits 150 tok/s Decode on a Single RTX 5090
A Qwen ambassador reports 150 tok/s decode for Qwen3.8 Flash-Next on one RTX 5090 using the Strata engine and IQ3_S quantization, with roughly 130 tok/s median…
-
Qwen Code Web Shell Can Now Be Installed as a Standalone App
A community pull request by jabrailkhalil lets Qwen Code Web Shell be installed from the browser as a standalone app with its own window and icon.
-
DeepSeek Open-Sources DeepInfer Inference Framework and Token Codec
DeepSeek released DeepInfer, its in-house inference framework that powers DeepSeek-V4.1-Flash at scale, along with deepseek-recipe, a token encoding and decodin…
-
DeepSeek Open-Sources DeepInfer Inference Framework Under Apache-2.0
DeepSeek released DeepInfer, a high-performance large-scale inference framework, under the Apache-2.0 license, with optimized kernels, MoE expert parallelism, d…
-
Tencent's Hy Memory Shrinks Local Footprint from 60GB to 18GB with llama.cpp
Tencent reports that its open-source Hy Memory stack can now run locally at 18GB instead of 60GB by replacing vLLM with a single llama.cpp binary, with quantize…
-
DeepSeek V4.1 Flash 1017 Offers $100 Usage Credit in Command Code for 24 Hours
Ahmad Awais says the latest DeepSeek V4.1 Flash 1017 build comes with $100 of usage in Command Code for the next 24 hours.
-
MiniMax M3.1 Flash Free on Ollama Cloud for 3 Days
Ollama and MiniMax are offering MiniMax M3.1 Flash free on Ollama Cloud for three days, hosted in the US with zero data retention.
