When will K3's weights be available, and what does 'open weights' mean?

K3's 2.8 trillion parameters will ship on July 27. Open weights means the model runs on private infrastructure, fine-tunes on private data, and cannot be switched off by the provider—unlike API models accessed through a vendor's infrastructure.

How much cheaper is K3's API than Claude Fable 5?

K3's API is priced at roughly a third of Fable 5's.

Which benchmarks does K3 beat Claude Opus 4.8 on?

K3 beats Opus 4.8 at max effort on four of five major benchmarks: BrowseComp, DeepSWE, GPQA Diamond, FrontierSWE, and HLE with tools.

AI Safety & AlignmentLarge Language ModelsDaily Dose of Data ScienceJul 18, 2026Jul 18, 2026, 10:00 JST2 min read

Moonshot's K3 beats Claude, GPT models; ships 2.8T open weights July 27

Send on LINE

Key takeaway

Moonshot AI released Kimi K3, a model that beats or matches top closed-source AI systems (Claude Opus 4.8, GPT-5.6 Sol) on major reasoning benchmarks while costing roughly a third of Claude Fable 5's API price.
The model is open weights with 2.8 trillion parameters and will ship fully on July 27, becoming the largest open model released to date—a significant alternative for users who want to avoid dependence on vendor-controlled APIs, though weights remain unavailable until that date.

3 Key Points

What happened
Moonshot AI released Kimi K3, which beats Claude Opus 4.8 on four of five major benchmarks (BrowseComp, DeepSWE, GPQA Diamond, FrontierSWE, HLE with tools) and matches or exceeds Claude Fable 5 and GPT-5.6 Sol on several others. K3 is open weights with 2.8 trillion parameters, scheduled to ship on July 27, making it the largest open model released so far. Its API is priced at roughly a third of Fable 5's.
Why it matters
Open-weight models run on private infrastructure and cannot be shut down by a provider—a meaningful difference after Anthropic was forced to take Fable 5 offline globally for 18 days due to a US export directive last month. Open weights also allow fine-tuning on private data. However, K3's weights are not yet public; until July 27 it functions as a standard API model like any other.
What to watch
K3 ships July 27. Until then, access is API-only. The model's actual performance on the announced benchmarks and real-world usage will determine whether it sustains its claimed advantages over closed frontier models.

Ask the AI about this article →

Summaries like this, in your inbox every morning.

Get news like this every morning — free

Context & Analysis

The release of Kimi K3 reflects a strategic shift in the competitive dynamics of large language models. Moonshot AI has created a model that performs at or above the level of leading closed-source systems—Claude Opus 4.8, Claude Fable 5, and GPT-5.6 Sol—while simultaneously offering a significantly lower API price and, crucially, an open-weight variant. This positioning exploits a growing concern among enterprises: the risk of vendor lock-in and service interruption. Anthropic's recent 18-day global outage of Fable 5, forced by a US export directive, underscores the vulnerability of closed models accessed exclusively through vendor-controlled APIs. Open weights eliminate that failure mode by running on private infrastructure and allowing users to retain full control.

The scale of K3 (2.8 trillion parameters) matters because it will be the largest open model released to date, broadening the audience of organizations that can deploy frontier-class capability without API dependency. However, a critical caveat applies: until July 27, K3 functions as a standard API model, so early adoption still carries the same operational risk as closed alternatives. The pricing advantage—roughly one-third of Fable 5's—combined with the promise of open weights creates a credible alternative for cost-sensitive and autonomy-focused buyers, potentially reshaping how enterprises evaluate frontier models.

FAQ

When will K3's weights be available, and what does 'open weights' mean?: K3's 2.8 trillion parameters will ship on July 27. Open weights means the model runs on private infrastructure, fine-tunes on private data, and cannot be switched off by the provider—unlike API models accessed through a vendor's infrastructure.
How much cheaper is K3's API than Claude Fable 5?: K3's API is priced at roughly a third of Fable 5's.
Which benchmarks does K3 beat Claude Opus 4.8 on?: K3 beats Opus 4.8 at max effort on four of five major benchmarks: BrowseComp, DeepSWE, GPQA Diamond, FrontierSWE, and HLE with tools.

Daily Dose of Data ScienceRead Original Article

Send on LINE