NewsHive
CONTACT USANALYST PORTAL →
BACKGROUND

ByteDance is at an early stage of training a model with as many as 10 trillion p

BACKGROUND
1 SIGNALFIRST DETECTED 7 August 2026
Narrative synthesis pending — 1 signal collected.
How the story developed
Other feeds and industry voices
8 Jul
Minimax plans to release a 2.7-trillion parameter model.

<table> <tr><td> <a href="https://www.reddit.com/r/singularity/comments/1uqueil/minimax_plans_to_release_a_27trillion_parameter/"> <img alt="Minimax plans to release a

OPINION · WHAT · singularityOPEN STORY
7 Aug
ByteDance trains massive AI model in bid to rival Anthropic

TikTok owner training a model with 10 trillion parameters.

REPORT · WHAT · Ars TechnicaOPEN STORY
19 May
Cerebras is running a trillion parameter model (Kimi K2.6) at 1000 tokens/s

<table> <tr><td> <a href="https://www.reddit.com/r/singularity/comments/1thw41i/cerebras_is_running_a_trillion_parameter_model/"> <img alt="Cerebras is running a trillion

OPINION · WHAT · singularityOPEN STORY
1 Oct
Google releasing their most powerful model yet

<table> <tr><td> <a href="https://www.reddit.com/r/singularity/comments/1wv0jvd/google_releasing_their_most_powerful_model_yet/"> <img alt="Google releasing their most powerful

OPINION · WHAT · singularityOPEN STORY
19 May
bytedance released an open source model that attempts to do just about anything with only 3b parameters

<table> <tr><td> <a href="https://www.reddit.com/r/LocalLLaMA/comments/1thkwgk/bytedance_released_an_open_source_model_that/"> <img alt="bytedance released an open source model

OPINION · WHAT · LocalLLaMAOPEN STORY
10 Aug
ByteDance vows to avoid AI distillation, develop new model its own way

<table> <tr><td> <a href="https://www.reddit.com/r/LocalLLaMA/comments/1vk7o93/bytedance_vows_to_avoid_ai_distillation_develop/"> <img alt="ByteDance vows to avoid AI

OPINION · WHAT · LocalLLaMAOPEN STORY
10 Aug
The Chinese tech giant is at an early stage of training a model with as many as 10 trillion parameters—three times larger than Moonshot’s Kimi K3, the biggest Chinese model released to date.

The Chinese tech giant is at an early stage of training a model with as many as 10 trillion parameters—three times larger than Moonshot’s Kimi K3, the biggest Chinese model

OPINION · WHAT · Ars TechnicaOPEN STORY
7 Jul
SpaceXAI planning to launch 1.5 Trillion parameter Grok 4.5 on Wednesday

<table> <tr><td> <a href="https://www.reddit.com/r/singularity/comments/1uq9qjq/spacexai_planning_to_launch_15_trillion_parameter/"> <img alt="SpaceXAI planning to launch 1.5

OPINION · WHAT · singularityOPEN STORY
Sources
singularity

NewsHive monitors these sources continuously. All signal titles above link to the original reporting.

Intelligence by NewsHive. Need help navigating what this means for your business? Contact GeekyBee →