Skip to content

Alibaba News: Latest Releases & Events

In October 2026, OpenAI made its fourth consecutive price cut amid pressure from open-weight Chinese providers including Alibaba. In August, Alibaba's Qwen 3.8 27B matched GPT-5.6 Luna (max) on the Artificial Analysis Intelligence Index, after the 2.4-trillion-parameter Qwen3.8-Max launched. In March, Alibaba Cloud raised prices by up to 34%, citing rising hardware costs and surging global AI demand.

Who is Alibaba? → Big AI Players

All 8 Alibaba entries of the AI News timeline, newest first.

2026

October 2026

2026-10-26

  • OpenAI releases GPT-6.1 Terra at half price
    OpenAI introduced GPT-6.1 Terra, a fast, efficient model for computer-based tasks like drafting proposals and editing video, priced at $1 per million input tokens and $5 per million output tokens — 50% cheaper than GPT-6.1 Sol and the fourth consecutive OpenAI price cut amid pressure from open-weight Chinese providers like Alibaba, Baidu, and Z.ai. Analysts note the escalating price war gives enterprises lower-cost options but deepens confusion over which model fits which workload, with implications for both OpenAI and Anthropic as they pursue public listings. (aibusiness.com)
    OpenAI Baidu Zhipu AI

August 2026

2026-08-17

  • Qwen 3.8 27B scores 52 on Intelligence Index
    Qwen 3.8 27B scored 52 on the Artificial Analysis Intelligence Index, matching GPT-5.6 Luna (max) and falling just one point behind GLM-5.2 (753B parameters) and DeepSeek V4 Pro 0813 (1.7T parameters), despite being far smaller than any of those models. Simon Willison calls it "a truly astonishing model," highlighting the efficiency of matching trillion-parameter competitors at only 27B parameters. (simonwillison.net)

2026-08-14

  • Alibaba releases Qwen 3.8 27B
    Alibaba's Qwen lab released Qwen 3.8 27B, an Apache 2.0-licensed, vision-capable 27B parameter open-weights LLM whose self-reported benchmarks beat both its predecessor and the closed-weight Qwen 3.7-Plus. In hands-on testing, Simon Willison found it capable of strong code generation, tool calling, coding-agent loops, and image bounding-box tasks on consumer hardware (a 17GB Q4_K_M quantization on a MacBook Pro and NVIDIA DGX Spark), but warned that its default 'xhigh' reasoning effort causes wildly excessive overthinking (e.g., 22,276 reasoning tokens for an SVG drawing) and recommended running it at low or no reasoning, while noting inference is slow at ~15-30 tokens/second though Multi-Token Prediction via llama.cpp improved throughput by roughly 72%. (simonwillison.net)

2026-08-03

  • Alibaba Unveils Qwen3.8-Max AI Model
    Alibaba introduced Qwen3.8-Max, its largest and most powerful open-weight AI model to date, featuring 2.4 trillion parameters and a context window of up to 1 million tokens. The model is designed to handle complex, long-horizon workloads including autonomous coding, legal document review, and financial research. (aibusiness.com)

July 2026

2026-07-19

  • Alibaba Previews Qwen 3.8 Max Model
    Alibaba previewed its new flagship AI model, Qwen 3.8 Max, featuring 2.4 trillion parameters and claiming performance that ranks just behind Anthropic's Fable 5. The release underscores the growing competitiveness of cost-effective, open-weight Chinese AI models against leading U.S. providers. (aibusiness.com)

March 2026

2026-03-18

  • Alibaba Cloud hikes prices up to 34%
    Alibaba Cloud increased prices for compute, storage, and SaaS services by up to 34%, citing rising hardware costs and surging global AI demand. The adjustments affect various instance types, including high-end GPU instances and Alibaba's own silicon, with existing prices honored until renewal cycles begin after April 18, 2026. (go.theregister.com)

February 2026

2026-02-16

  • Qwen Team Announces Qwen 3.5 Open Weights Model
    The Qwen team announced the release of Qwen 3.5, marking the latest update to their large language model series. Called Qwen3.5-397B-A17B it features 397 billion total parameters with 17 billion activated per forward pass. It is a native vision-language model, supporting 201 languages. (qwen.ai)

2026-02-10