Kimi K2.7-Code cuts thinking tokens 30% — but practitioners say the benchmarks don't check out

Kimi K2.7-Code claims 30% fewer thinking tokens and a drop-in API swap path, but independent benchmarks show kernel regressions and no DeepSWE submission.
Forbes

Chinese AI Model Developer Kimi Raising Funds Valuing It At $20 Billion

This voice experience is generated by AI. Learn more. This voice experience is generated by AI. Learn more. Kimi founder Yang Zhilin. Kimi, the Chinese AI model startup also known as Moonshot AI, is ...
Geeky Gadgets

Kimi K2 Thinking Beats GPT-5 : Meet the AI Model You Can Run for Cents

What if the most powerful AI model wasn’t locked behind corporate paywalls or billion-dollar budgets? Imagine an open source tool so advanced it could rival proprietary giants like GPT-5 and Sonnet ...
Geeky Gadgets

Kimi K2 Agent Researcher for Deep Reasoning Research Tasks

What if you could delegate your most complex research tasks to an AI that not only understands the intricacies of your work but also evolves with every challenge it faces? Enter the Kimi K2 Agent ...
moneycontrol.com

Moonshot AI releases Kimi 2.5 model that claims to outperform LLMs from OpenAI, Google

Chinese companies are accelerating the rollout of new artificial intelligence (AI) models as competition against US-based rivals such as OpenAI, Anthropic and Google escalates. While the US companies ...