RSS.bz

AI Daily · Oct 27, 2026: OpenAI launches GPT-6.1 Terra: stronger computer use at $1/M input tokens

GPT-6.1 Terra ships via API at $1/M input and $5/M output with stronger writing and cross-app computer use, matching Astra on Agents' Last Exam at one-tenth the cost. Meanwhile Qoder makes DeepSeek V4 models free for a limited time.

Pick a date 10/2026: 27 briefings
10/2026
SunMonTueWedThuFriSat
123
45678910
11121314151617
18192021222324
25262728293031

Dotted dates have a briefing; click one to open it.

  • OpenAI releases GPT-6.1 Terra with stronger writing and computer use

    GPT-6.1 Terra is OpenAI's first model notably better at writing, personality and nuanced request understanding, with strong computer use and 2x the usage of 6.1 Sol across API and subscriptions. It is priced at $1/M input and $5/M output with 1.05M context, matching GPT-6 Astra on Agents' Last Exam at one-tenth the per-token price and finishing tasks over 3x faster.

    @thsottiaux 9 duplicate reports merged

  • DeepSeek-V4.1-Flash-1017 free on Qoder until Nov 1

    Qoder is offering DeepSeek-V4.1-Flash-1017 free to all users, including free-tier accounts, until Nov 1 00:00 UTC+8; afterwards usage reverts to 0.1x credit cost.

    @qoder_ai_ide 4 duplicate reports merged

  • DeepSeek-V4-Pro-0813 free in Qoder through October 31

    DeepSeek-V4-Pro-0813 is free in Qoder through Oct 31 for both free and paid individual accounts, with daily free uses costing 0 credits even at zero balance and no claim required.

    @qoder_ai_ide

  • Qoder offers 12,288 SOTA Credits over three days to Pro subscribers

    Qoder Global Pro and above individual subscribers, plus approved Community Boost members, can claim 4,096 SOTA Credits per day on Oct 28, 29 and 30, each day's pack opening 00:00-23:59 UTC+8 with no carryover.

    @qoder_ai_ide

  • DeepSeek V4.1 Flash added to Cursor at $0.60–$0.98 per task

    Cursor now offers DeepSeek V4.1 Flash at every effort level, with per-task cost on CursorBench ranging from $0.60 to $0.98 and no model scoring higher on that cost frontier.

    @cursor_ai

  • MiniMax M3.1 Flash runs locally via Unsloth 2-bit GGUF at 143.9GB

    Unsloth's Dynamic GGUF brings the 428B MoE MiniMax M3.1 Flash to a 143.9GB 2-bit footprint with 91.5% top-1 token agreement across seven domains, runnable via Unsloth Desktop, llama.cpp, and an OpenAI-compatible API.

    @MiniMax_AI

  • Cursor launches Agent SDK for defining and deploying agents as code

    Cursor's Agent SDK lets developers define agents as code with prompts in Markdown and tools in TypeScript, deploy them on Cursor's hosted infrastructure or self-host, and use built-in evals to measure changes over time.

    @cursor_ai 1 duplicate reports merged

  • Qoder Lite plan costs $10/month for 1,000 credits

    Qoder Lite is priced at $10/month with 1,000 credits; new subscribers get 2,000 credits in the first month and daily check-in adds 100 credits. Qoder cites tests with Qwen3.8-Flash covering 434 PDFs, 125 code reviews or 200 test suites.

    @qoder_ai_ide

  • GMI Cloud offers 15% off MiMo V2.6 Flash until November 1

    GMI Cloud is running a limited-time 15% discount on MiMo V2.6 Flash inference, ending November 1.

    @gmi_cloud 1 duplicate reports merged

  • 1,300+ 3D models open-sourced on Hugging Face, 39 GB free

    Over 1,300 3D models totaling 39 GB are free to download on Hugging Face, including input images, prompts and settings, GLBs with up to 8K textures and native Blender files; generated with Hitem3D, Meshy, Rodin, Tripo, Claude Opus 5.5 3D and GPT-6 Astra 3D on Scenario.

    @huggingface