Kimi K3 scores 57 on the Artificial Analysis Intelligence Index. Its intelligence is comparable to Opus 4.8 and GPT-5.5 but remains behind Fable 5 and GPT-5.6 Sol. Moonshot AI has expressed plans to release the 2.8T parameter model's weights, which would make it the leading open weights model Key r
Moonshot AI's Kimi K3 model achieves strong benchmark scores and plans to release its 2.8T parameter weights, positioning it as the leading open weights model.
The fuller picture from Moonshot's own launch post: K3 is a 2.8T-parameter monster, the first open model at that scale, and by their own admission it still trails Fable 5 overall. But it beats Opus 4.8 on much of their benchmark table, and the weights go public July 27. An open model trading blows
Kimi K3 scores 57 on the Artificial Analysis Intelligence Index. Its intelligence is comparable to Opus 4.8 and GPT-5.5 but remains behind Fable 5 and GPT-5.6 Sol. Moonshot AI has expressed plans to release the 2.8T parameter model's weights, which would make it the leading open weights model Key r
There’s a new model in the arena… Kimi-K3 is currently ranking higher than Anthropic’s Fable 5 for front end web development tasks $3 in / $15 out vs $10 in / $50 out Moonshot nearly 4x’d their own pricing with K3 and it’s still 30% of Fable’s cost… https://t.co/QXO3JSTBCt
Introducing Kimi K3: Open Frontier Intelligence 🔹 2.8 Trillion Parameters, 1 Million Context, Native Multimodal 🔹 Kimi Delta Attention enables up to 6.3x faster decoding in million-token contexts 🔹 Attention Residuals deliver ~25% higher training efficiency at <2% additional cost 🔹 Built for lo
Kimi K3 scores 57 on the Artificial Analysis Intelligence Index. Its intelligence is comparable to Opus 4.8 and GPT-5.5 but remains behind Fable 5 and GPT-5.6 Sol. Moonshot AI has expressed plans to release the 2.8T parameter model's weights, which would make it the leading open weights model Key r
Kimi K3 scores 57 on the Artificial Analysis Intelligence Index. Its intelligence is comparable to Opus 4.8 and GPT-5.5 but remains behind Fable 5 and GPT-5.6 Sol. Moonshot AI has expressed plans to release the 2.8T parameter model's weights, which would make it the leading open weights model Key r
RT @ArtificialAnlys: Kimi K3 scores 57 on the Artificial Analysis Intelligence Index. Its intelligence is comparable to Opus 4.8 and GPT-5.…
Moonshot AI unveils Kimi K3, the largest open-weight AI model at 2.7 trillion parameters
🐮月之暗面Kimi K3正式发布：2.8万亿参数，将在十天后开源 今日，月之暗面正式发布新一代旗舰模型Kimi K3。K3总参数达2.8万亿，采用混合专家架构。每次运行只激活896个专家中的16个。月之暗面称，这是首个3万亿级开放模型。 官方评测显示，K3在部分编程和Agent测试中超过 Claude Opus 4.8、GPT-5.5。整体仍落后Claude Fable 5和GPT-5.6 Sol。 K3已上线 Kimi、Kimi Work、Kimi Code和API。API输入价格为每百万token 3美元，输出为15美元。 完整权重计划最晚在7月27日发布。 此外，据月之暗面联