DAILY
AI daily picks · 2026-10-03
8 stories published here on 2026-10-03 (Beijing time), less than 48 hours after the event and above our selection bar, ranked by editorial score.
- 1Model releaseEvent time 1 independent sourceAleph Alpha releases Kolibri-1, an open-weight German-English reasoning model
Aleph Alpha released Kolibri-1 on October 3. The Apache 2.0 MoE model has about 78B total parameters and about 3.46B active per token, supports tool calling, and offers a maximum context of 1,048,576 tokens with a lower recommended operating length.
- 2Tutorial and hands-onEvent time 1 independent sourceOpenAI publishes a GPT-6 family model-selection guide
OpenAI published a guide for choosing among GPT-6 Astra, GPT-6.1 Sol and GPT-6 Luna. It connects task difficulty with model choice and recommends evaluating task success, latency and cost in production workflows.
- 3Tutorial and hands-onEvent time 1 independent sourceAWS documents multi-turn reinforcement learning for a search agent
AWS published a SageMaker AI workflow for training a Qwen3.6-27B search agent with multi-turn reinforcement learning. It uses retrieval tools and trajectory rewards, and reports gains on three of four evaluated held-out sets.
- 4Tutorial and hands-onEvent time 1 independent sourceAWS explains authenticated web search for Claude Desktop through AgentCore
AWS published a tutorial connecting Claude Desktop configured for Amazon Bedrock to AgentCore Gateway’s MCP web search. The example combines IAM Identity Center, Cognito and OAuth authentication with AWS-managed search.
- 5Model releaseEvent time 1 independent sourceAi2 releases AstaBrief, an open 8B model for cited research reports
Ai2 released AstaBrief, a Qwen3-8B-based open-weight model that turns a research question and retrieved literature excerpts into a cited report. It powers Asta Fast mode; Ai2 also publishes training data and an example PDF workflow.
- 6Product updateEvent time 1 independent sourceClaude Code 2.1.288 adds review controls and session recovery fixes
Anthropic released Claude Code 2.1.288 with a configurable code-review findings limit, recovery of drafts cleared by Ctrl+C, and fixes for timeout recovery in noninteractive and subagent responses.
- 7Product updateEvent time 1 independent sourceGitHub retires four older models across Copilot experiences
GitHub deprecated Gemini 3.5 Flash, Gemini 3.6 Flash, Kimi K2.7 Code, and Claude Opus 4.7 in Copilot on October 2. It recommends Gemini 3.8 Flash, Kimi K3, and Claude Opus 5.5 as corresponding alternatives.
- 8Product updateEvent time 1 independent sourceGitHub Copilot code review becomes available through REST and GraphQL
GitHub announced generally available REST and GraphQL support for requesting Copilot code reviews. The October 2 announcement also explains Balanced effort as the default, which took effect September 28; explicit Lite settings are preserved.