DeepSeek V4.1 Flash 1017 Ultrafast now available on AI Gateway
DeepSeek V4.1 Flash 1017 Ultrafast is available through Vercel AI Gateway, with a 1M-token context and generation above 1,000 tokens per second. Builders can try it in coding agents via the Vercel CLI.
Vercel AI Gateway now offers **DeepSeek V4.1 Flash 1017 Ultrafast** through Fireworks. It has a **1M-token context window** and can generate upwards of **1,000 tokens per second**; public preview has no waitroom or session time limit.
For interactive coding or agent workflows where waiting on each model step is a bottleneck, configure the model in Claude Code, Codex, Cursor, or another supported agent using the latest Vercel CLI.
Vercel AI Gateway now offers **DeepSeek V4.1 Flash 1017 Ultrafast** through Fireworks. It has a **1M-token context window** and can generate upwards of **1,000 tokens per second**; public preview has no waitroom or session time limit. For interactive coding or agent workflows where waiting on each model step is a bottleneck, configure the model in Claude Code, Codex, Cursor, or another supported agent using the latest Vercel CLI. The material gives no task-quality or latency comparison against other models. The preview’s speed and long context do not establish how well it performs on a particular codebase.
This adds a Vercel-routed DeepSeek option combining a 1M-token context window with claimed throughput above 1,000 tokens per second. Nearby listings already offer million-token coding routes, and one claims similar throughput; this launch expands evaluation choices but supplies no comparative quality or latency evidence.