Moonshot AI's Kimi K3 to launch with 2-3 trillion parameters, targets Anthropic Claude Opus 4.8 performance
Moonshot AI will release Kimi K3 in the coming days with a parameter count between 2 trillion and 3 trillion, according to Financial Times sources. The open-weight model is expected to perform at par with or surpass Anthropic's Claude Opus 4.8, making it the largest open-weight AI model from China.
Moonshot AI's Kimi K3 to launch with 2-3 trillion parameters, targets Anthropic Claude Opus 4.8 performance
Moonshot AI will release Kimi K3 in the coming days with a parameter count between 2 trillion and 3 trillion, according to anonymous sources cited by the Financial Times. The open-weight model is expected to perform at par with or potentially surpass Anthropic's Claude Opus 4.8.
Kimi K3 will be the largest open-weight AI model released from China to date. The company's previous Kimi K2 models have performed well on benchmarks and ranked high in the open-source AI market, though specific benchmark scores were not disclosed.
Funding and valuation
Moonshot AI is reportedly raising fresh capital in a round that would value the company at $31.5 billion. The company raised $2 billion at a $20 billion valuation in May 2025, marking a 57.5% increase in valuation over approximately 14 months.
Market context
The release comes as enterprises debate whether to pay premium prices for closed-source models from OpenAI and Anthropic versus using open-source alternatives. Industry executives have expressed concerns about data privacy with closed-source AI products like ChatGPT and Claude, where clients submit proprietary data for processing.
Open-source models from Chinese AI labs including DeepSeek, Z.ai, and Moonshot are increasingly positioned as alternatives that companies can train for specific use cases. The performance gap between these open models and frontier closed-source models has narrowed, according to industry sources.
What this means
If Kimi K3 delivers on claimed performance parity with Claude Opus 4.8, it would represent a significant shift in the open-weight model landscape. A 2-3 trillion parameter open-weight model matching frontier closed-source performance could accelerate enterprise adoption of self-hosted AI solutions, particularly among organizations prioritizing data sovereignty. However, actual benchmark scores and deployment costs remain undisclosed, making it difficult to assess the true competitive position against Anthropic's offering.
Related Articles
Z.ai Confirmed as Creator of Chart-Topping 'Ox Alpha' Model, Weights Coming Wednesday
Z.ai, maker of the GLM model series, has confirmed it is behind Ox Alpha, the mysterious open-weight model that appeared anonymously on OpenRouter and topped benchmark leaderboards. The company will release the model's weights on Wednesday.
Tencent Open-Sources Hy4 Preview: 770B-Parameter MoE Model with 1M-Token Context
Tencent's Hy Team has open-sourced Hy4 preview, a 770-billion-parameter Mixture-of-Experts model with 49 billion activated parameters and a 1-million-token context window. The model is available under Apache 2.0 alongside an FP8-quantized variant, with Tencent claiming it beats GLM 5.3 and Kimi K3 on internal engineering evaluations.
Tencent Releases Hy4 Preview: 770B-Parameter MoE Model with 1M Context for Coding Agents
Tencent has released Hy4 preview, a mixture-of-experts model with 770B total parameters and 49B active parameters, targeting coding agents and multi-step tool-use workflows. The model ships with a 1 million token context window and is priced at $0.834 per 1M input tokens and $2.501 per 1M output tokens.
Google Launches Gemini 3.5 Transcribe with 4.0% Word Error Rate Across 85 Languages
Google has released Gemini 3.5 Transcribe, a speech-to-text model that automatically detects 85 languages, removes filler words, and corrects misspoken phrases. The company claims a 4.0 percent word error rate for streaming audio and 70 percent lower latency than its predecessor, Chirp 3.
Comments
Loading...