Moonshot AI Releases Kimi K3, Claiming Benchmark Wins Over Claude Fable and GPT 5.6 Sol

Moonshot AI has introduced Kimi K3, a 2.8-trillion-parameter open-source model that the company says outperforms Claude Fable and GPT 5.6 Sol on selected benchmarks. The model is also being positioned at Claude Sonnet pricing.

Moonshot AI Releases Kimi K3, Claiming Benchmark Wins Over Claude Fable and GPT 5.6 Sol

What happened?

Moonshot AI has introduced Kimi K3, a 2.8-trillion-parameter open-source model that the company says outperforms Claude Fable and GPT 5.6 Sol on selected benchmarks. The model is also being positioned at Claude Sonnet pricing.

Why it matters

Moonshot AI has launched Kimi K3, a 2.8-trillion-parameter open-source AI model that the company says beats Claude Fable and GPT 5.6 Sol on key benchmarks. According to the source material, the model leads a creative writing benchmark and also tops Arena AI’s frontend code leaderboard.

Moonshot AI has launched Kimi K3, a 2.8-trillion-parameter open-source AI model that the company says beats Claude Fable and GPT 5.6 Sol on key benchmarks. According to the source material, the model leads a creative writing benchmark and also tops Arena AI’s frontend code leaderboard.

The release matters because it adds another large open-source model to the fast-moving AI race, where benchmark performance and pricing help shape how developers and companies choose tools. For readers in crypto and adjacent tech markets, developments like this can influence broader sentiment around infrastructure, compute demand, and the competition among AI platforms.

Moonshot AI is presenting Kimi K3 as a high-capacity model with performance claims that place it alongside top-tier proprietary systems. The source also says the model is being offered at Claude Sonnet pricing, which may be relevant for teams comparing cost against capability.

The benchmark results highlighted in the announcement focus on creative writing and frontend coding. As with all model comparisons, the practical value depends on the task, the evaluation method, and how the model performs in real-world use beyond leaderboard settings.

Source: Decrypt

Keep exploring

Related stories

Bitget to Leave Japan and Close Remaining Positions by Year-End

Bitget to Leave Japan and Close Remaining Positions by Year-End

Bitget plans to exit the Japanese market and close all remaining positions by the end of the year. The move adds to ongoing shifts among crypto companies navigating Japan’s regulatory environment.

Read
State of Crypto countdown highlights key industry developments

State of Crypto countdown highlights key industry developments

CoinDesk’s State of Crypto coverage counts down recent developments shaping the crypto policy and market landscape. The report frames these changes as important for how companies and participants navigate the sector.

Read
Strategy Keeps STRC Dividend at 12%

Strategy Keeps STRC Dividend at 12%

Strategy said it is holding the dividend on its STRC preferred stock at 12%. The decision keeps the payout unchanged for investors in the company’s yield-focused security.

Read