model release

Moonshot AI Launches 'Kimi Latest' Router Model with 262K Context Window

TL;DR

Moonshot AI released Kimi Latest, a router endpoint that automatically redirects to the most recent model in the Kimi family. The model features a 262,144 token context window, though specific pricing and performance benchmarks have not been disclosed.

2 min read
0

Kimi Latest — Quick Specs

Context window262K tokens
Input$3/1M tokens
Output$15/1M tokens

Moonshot AI Launches 'Kimi Latest' Router Model with 262K Context Window

Moonshot AI released Kimi Latest on April 27, 2026, a router endpoint that automatically redirects requests to the most recent model in the company's Kimi family. The model offers a 262,144 token context window.

Technical Specifications

The router model is designed to always point to the latest version of Moonshot AI's Kimi models, eliminating the need for developers to manually update model endpoints when new versions are released. This approach mirrors practices used by other AI providers for maintaining current model access.

The 262,144 token context window (approximately 262K tokens) positions the model among large context offerings in the market, though Moonshot AI has not disclosed whether this represents an increase from previous Kimi models.

Availability and Pricing

The model is available through OpenRouter, which normalizes API requests across providers. Pricing per million tokens has not been disclosed. OpenRouter indicates the model supports reasoning-enabled features, allowing access to step-by-step thinking processes through the reasoning_details array in API responses.

No usage activity data is currently available on OpenRouter's platform, as indicated by the provider's dashboard.

Integration Details

Developers can access Kimi Latest through OpenRouter's API using the model identifier ~moonshotai/kimi-latest. The platform supports both OpenRouter SDK and OpenAI SDK integration methods, along with various third-party frameworks.

What This Means

Router endpoints that auto-update to the latest model version reduce integration maintenance for developers but sacrifice version pinning control. The 262K context window suggests Moonshot AI is competing in the long-context segment, though without disclosed pricing or benchmark scores, it's difficult to assess competitiveness against established models from Anthropic (Claude 3.5 with 200K context) or Google (Gemini 1.5 Pro with 2M context). The lack of usage data and performance metrics on OpenRouter may indicate a very recent release or limited early adoption.

Related Articles

model release

DeepSeek V4.1-Flash Cuts KV Cache Memory by Up to 8x, Matches Opus 5 on Coding Benchmark

DeepSeek released V4.1-Flash, a 552-billion-parameter model built to slash the memory overhead of long-context AI agents. The model cuts GPU cache needs to roughly a quarter of its predecessor's and matches closed models from OpenAI and Anthropic on select coding benchmarks.

model release

Alibaba's Qwen Releases Qwen-Drive-1.0-4B, a Unified VLM for Autonomous Driving Perception and Planning

Alibaba's Qwen team has released Qwen-Drive-1.0-4B, a 4B-parameter vision-language model built on Qwen3.5 that unifies 3D perception, driving question answering, and motion planning in one framework. The model reports strong open-loop, pseudo-closed-loop, and closed-loop driving benchmark results while claiming minimal loss of general vision-language ability.

model release

DeepSeek Launches V4.1 Flash: Low-Cost MoE Model Claims to Beat V4 Pro

DeepSeek has released V4.1 Flash, a sparse mixture-of-experts model priced at $0.30 per 1M input tokens and $1.20 per 1M output tokens with a 1 million token context window. DeepSeek claims the model exceeds the larger V4 Pro on performance, speed, and task completion time.

model release

DeepSeek Releases V4.1-Flash: 552B MoE Model Cuts KV Cache to 890 Bytes Per Token

DeepSeek has released V4.1-Flash, a 552B-parameter multimodal Mixture-of-Experts model supporting 1M-token context and activating only 8B parameters during prefill. The model uses a new Causal Encoder-Decoder architecture and Compressed Sparse Attention 2 to cut global KV cache to 890 bytes per token, roughly a quarter of its predecessor.

Comments

Loading...