AI industry competition intensifies with new model releases from Alibaba, SpaceXAI, and others

Artificial Intelligence Large Language Models Open-source AI Qwen3.8 Grok 4.6

The artificial intelligence sector is experiencing a period of intense competition as several major companies release new large language models. Alibaba has released the open weights for its Qwen3.8-2.4T-A95B model, a mixture-of-experts architecture featuring 2.4 trillion total parameters and 95 billion active parameters. Additionally, Alibaba introduced the Qwen3.8-27B model, a lightweight version optimized for local deployment on personal hardware.

SpaceXAI has launched Grok 4.6, which has achieved a 61 score on the Artificial Analysis Intelligence Index, placing it on par with OpenAI's GPT-5.6 Sol. Grok 4.6 is positioned as a cost-effective option, priced at $2 per million input tokens and $6 per million output tokens. Meanwhile, DeepSeek released its V4 Pro model, though the company has faced scrutiny following announcements of increased API fees during peak hours.

Other significant updates include Google's release of Gemini 3.7 Flash, which emphasizes speed with an output of 340 tokens per second, and OpenAI's demonstration of an 'UltraFast' mode for GPT-5.6 Sol, capable of reaching 750 tokens per second through a partnership with Cerebras. These simultaneous releases signal a market shift where advanced AI capabilities are increasingly competing on the basis of price, openness, and specialized performance.

Entities

Alibaba · Anthropic · DeepSeek · Google · OpenAI · Qwen · SpaceXAI

Claims

What the coverage asserts, and how well corroborated each claim is across sources.

[● 2 SOURCES] The Qwen3.8-2.4T-A95B model uses a mixture-of-experts design with 2.4 trillion total parameters and 95 billion active parameters. nordictimes.com · nyadagbladet.se
[● 5 SOURCES] The Artificial Analysis Intelligence Index scores Claude Opus 5 at 63 and Grok 4.6 at 61. ai.cnmo.com · www.techritual.com · www.aktuelno.me · cryptobriefing.com · memeburn.com
[● 2 SOURCES] Alibaba released the weights for its Qwen3.8-2.4T-A95B model openly on Hugging Face and ModelScope. nordictimes.com · nyadagbladet.se
[○ 1 SOURCE] Claude Opus 5 features a 1 million token context window.
[● 5 SOURCES] SpaceXAI launched Grok 4.6 with an input price of $2 per million tokens and an output price of $6 per million tokens. ai.cnmo.com · www.techritual.com · www.aktuelno.me · cryptobriefing.com · memeburn.com
[● 2 SOURCES] DeepSeek announced an increase in API fees, with cache hit input fees increasing nearly 12-fold during peak hours. ai.cnmo.com · www.techritual.com
[● 2 SOURCES] Google released Gemini 3.7 Flash, which features an output speed of 340 tokens per second. ai.cnmo.com · www.techritual.com
[○ 1 SOURCE] Alibaba unveiled its compact Qwen3.8-27B model, designed for local laptop deployment. newsable.asianetnews.com
[● 2 SOURCES] OpenAI demonstrated an 'UltraFast' mode for GPT-5.6 Sol, achieving up to 750 tokens per second through a partnership with Cerebras. ai.cnmo.com · www.techritual.com