Kimi K2.7 Code High Speed

Kimi K2.7 Code High Speed is presented as a high-speed, coding-focused variant within the broader Kimi K2.7 family, designed to balance long-context understanding with the throughput needed for interactive developer workflows. It accepts text along with image and PDF inputs while producing text outputs, and it supports reasoning, tool calling, structured output, and temperature control, making it well suited to agent-style loops where the model must plan, invoke external tools, and continue generating coherent code in successive turns.

Beyond raw generation, the model is positioned for practical engineering tasks rather than as a general-purpose conversational system. Its large context window and substantial output budget allow it to ingest sizable codebases or documentation while still returning long, structured responses such as refactors, multi-file patches, or detailed explanations. Fast output token rates and low cache-read pricing further reduce the cost of iterative coding sessions, and the combined vision and PDF support lets teams route diagram-, screenshot-, and document-rich tasks through the same model without splitting pipelines.

Vercel AI Gatewaymoonshotai/kimi-k2.7-code-highspeedkimi-k2

Open provider page Provider docs

Quick Info

Provider: Vercel AI Gateway
Model key: moonshotai/kimi-k2.7-code-highspeed
Release date: Jun 12, 2026
Last updated: Jun 12, 2026
Knowledge cutoff: 2025-01
Input modalities
Output modalities
Capabilities

Cost

Input token cost: $1.90
Output token cost: $8.00

Limits

Output tokens: 32,768 tokens
Context window: 262,144 tokens

Transparent token rates

Compare Kimi K2.7 Code High Speed pricing

Rates are shown per one million tokens. Combined means one million input plus one million output tokens.

Browse this family

Latest news about Kimi K2.7 Code High Speed

Vercel AI Gateway

Coverage

Moonshot AI announced on June 15, 2026 the official launch of the Kimi K2.7 Code high-speed version, available initially to Kimi Code Beta program members, API developers, and Kimi Business users. The model logic remains consistent with the standard Kimi K2.7 Code released three days earlier on June 12, 2026, but the h In practical programming workloads the high-speed variant sustains roughly 180 tokens per second at median input lengths and peaks at up to 260 tokens per second on short-context tasks, and users must enable thinking mode to realize these speeds. Pricing is set at twice the standard Kimi K2.7 Code rate, specifically 13

Vercel AI Gateway

Coverage

Moonshot AI ("Dark Side of the Moon") officially launched the Kimi K2.7 Code Model High-Speed Edition on June 15, 2026, as an optimized variant of its previously released Kimi K2.7 Code programming model. The High-Speed Edition targets developers and enterprises needing frequent code iteration and rapid prototype valid According to the Baidu Baike entry, output speed reaches approximately 180 tokens per second in regular programming scenarios and up to 260 tokens per second in short-context scenarios, making the High-Speed Edition roughly 5 to 6 times faster than the standard Kimi K2.7 Code. At launch the variant was made available t