Skip to main content
LIVE SUN, 6 SEPT, 2026 BENGALURU · 28°C EDITION № 129 · FREE · NO LOGIN
AI AI · 1 MIN READ

Qwen 3.8 27B model now available on Cerebras at 1500 tokens per second

The Qwen 3.8 27B language model is now accessible on Cerebras public endpoints, delivering a processing speed of 1500 tokens per second.

The Qwen 3.8 27B language model is now accessible on Cerebras public endpoints, delivering a processing speed of 1500 tokens per second. This availability applies to both free trial and pay-as-you-go tiers, allowing developers to integrate the model into their applications starting immediately, according to Cerebras documentation.

Users can access Qwen 3.8 27B through Cerebras’ API, which supports multiple tiers including free trial and pay-as-you-go, with rate limits and pricing detailed on the platform. For those requiring higher throughput or dedicated service level agreements, Cerebras offers reserved capacity and dedicated endpoints. The company provides a quickstart guide and a model selection guide to help users choose the appropriate model for their use case.

The deployment of Qwen 3.8 27B on Cerebras’ infrastructure underscores the growing trend of making large language models more accessible and scalable for various applications. With a throughput of 1500 tokens per second, this model competes with other large-scale AI offerings, enabling faster inference times and potentially enhancing user experience in natural language processing tasks.

Developers interested in using Qwen 3.8 27B can begin by visiting Cerebras’ model catalog and following the quickstart instructions. The availability of this model on public endpoints marks a significant step in broadening access to advanced AI capabilities.

Editorial standards. Reported and edited at Startupniti's news desk from the source listed in the right rail. Every fact traces to a citation. If something looks wrong, write to corrections.
▸ WIRE
Premium content free for first 12 months · sign up to unlock Razorpay subscriptions launch Jan 2027 — ₹199/mo or ₹999/yr Every story reads every Indian tech source so you don't have to Every article cited · trust the source, not just the byline India's startup desk, edited daily Founders · Funding · Policy · Tech — three crawls a day Premium content free for first 12 months · sign up to unlock Razorpay subscriptions launch Jan 2027 — ₹199/mo or ₹999/yr Every story reads every Indian tech source so you don't have to Every article cited · trust the source, not just the byline India's startup desk, edited daily Founders · Funding · Policy · Tech — three crawls a day