# Boundless

> Boundless is the inference partner for AI-native companies, fitting the model and how it's served to your workload so each token costs less.

Boundless (boundless.network) is an AI infrastructure company. Boundless engineers identify the right open-weight model for a customer's workload, tune the serving stack to their traffic, and operate it in production. Customers are AI-native companies whose product depends on inference.

## Pages

- [Home](https://boundless.network/): positioning, approach, and how the engagement works
- [Forward-deployed inference](https://boundless.network/product/forward-deployed-inference): the engagement where Boundless inference engineers evaluate models, build the serving system, and operate it in production for a customer workload (model selection and evals, serving-system design, benchmarking and deployment, operations and changes)
- [Model library](https://boundless.network/models): every model Boundless serves, open-weight and frontier, with context windows and current pricing
- Model pages (https://boundless.network/models/<id>, markdown at /models/<id>.md): one per open-weight model, with prices, the same model at other providers, context window, capabilities, a runnable example and an FAQ
- [Playground](https://inference.boundless.network/playground): try any supported model, one free request
- [News](https://boundless.network/news): announcements and writing from the Boundless team (RSS: https://boundless.network/news/feed.xml)

## News

- [Laya is now available on Vercel AI Gateway, free through Boundless](https://boundless.network/news/laya-vercel-ai-gateway) (2026-09-30): Laya is now on Vercel AI Gateway, served by Boundless and free for a limited time. Explore structured decisions for agents and workflows.
- [Introducing Boundless AI, the inference partner for AI-native startups](https://boundless.network/news/introducing-boundless-ai-the-inference-partner-for-ai-native-startups) (2026-09-03): Boundless AI now in early access. We identify the right open-weight model for your workload, tune the serving stack, and operate it in production.

## Models

- [DeepSeek V4 Flash](https://boundless.network/models/dsv4) (DeepSeek): Low-cost text agents with long prompts and reusable prefixes. 1M context; $0.04 input, $0.02 cached input, $0.21 output per 1M tokens. Markdown: https://boundless.network/models/dsv4.md
- [MiMo-V2.6 Flash](https://boundless.network/models/mimo-v2.6-flash) (Xiaomi): Low-cost omnimodal work across text, images, audio, and video. 1M context; $0.14 input, $0.003 cached input, $0.28 output per 1M tokens. Markdown: https://boundless.network/models/mimo-v2.6-flash.md
- [Qwen3.6-35B-A3B](https://boundless.network/models/qwen3.6) (Qwen): Affordable image and text tasks with a shorter context. 262K context; $0.04 input, $0.02 cached input, $0.35 output per 1M tokens. Markdown: https://boundless.network/models/qwen3.6.md
- [GLM-5.3 Flash](https://boundless.network/models/glm-5.3-flash) (Z.ai): Lower-cost coding and visual tasks with a million-token context. 1M context; $0.12 input, $0.02 cached input, $0.40 output per 1M tokens. Markdown: https://boundless.network/models/glm-5.3-flash.md
- [Qwen3.8 Flash](https://boundless.network/models/qwen3.8-flash) (Qwen): Quick image analysis with a million-token context. 1M context; $0.15 input, $0.02 cached input, $0.47 output per 1M tokens. Markdown: https://boundless.network/models/qwen3.8-flash.md
- [GPT-6 Luna](https://inference.boundless.network/models/gpt-6-luna) (OpenAI): Everyday GPT-6 work where you want to keep costs down. 1M context; $0.10 input, $0.01 cached input, $0.50 output per 1M tokens.
- [Nemotron 3 Super](https://boundless.network/models/nemotron3-super) (NVIDIA): Coordinating agents, planning work, and combining their results. 262K context; $0.12 input, $0.06 cached input, $0.52 output per 1M tokens. Markdown: https://boundless.network/models/nemotron3-super.md
- [MiMo-V2.6 Pro](https://boundless.network/models/mimo-v2.6-pro) (Xiaomi): A frontier open-weight agent that reads text, images, audio, and video. 1M context; $0.44 input, $0.004 cached input, $0.87 output per 1M tokens. Markdown: https://boundless.network/models/mimo-v2.6-pro.md
- [DeepSeek V4.1 Flash](https://boundless.network/models/deepseek-v4.1-flash) (DeepSeek): Fast coding agents that need image input and a long context. 1M context; $0.20 input, $0.01 cached input, $1.00 output per 1M tokens. Markdown: https://boundless.network/models/deepseek-v4.1-flash.md
- [MiniMax M3](https://boundless.network/models/minimax-m3) (MiniMax): Long-running coding agents that work with images and video. 1M context; $0.28 input, $0.06 cached input, $1.10 output per 1M tokens. Markdown: https://boundless.network/models/minimax-m3.md
- [GPT-5.6 Luna](https://inference.boundless.network/models/gpt-5.6-luna) (OpenAI): High-volume GPT-5.6 chat, classification, and light agents. 262K context; $0.20 input, $0.02 cached input, $1.20 output per 1M tokens.
- [DeepSeek V4 Pro](https://boundless.network/models/deepseek-v4-pro-0813) (DeepSeek): A main coding agent paired with DeepSeek V4 Flash sub-agents. 1M context; $1.04 input, $0.04 cached input, $2.08 output per 1M tokens. Markdown: https://boundless.network/models/deepseek-v4-pro-0813.md
- [GLM-5.2](https://boundless.network/models/glm-5.2) (Z.ai): Long coding tasks that run for hours and work with text. 1M context; $0.34 input, $0.17 cached input, $2.30 output per 1M tokens. Markdown: https://boundless.network/models/glm-5.2.md
- [GLM-5.3](https://boundless.network/models/glm-5.3) (Z.ai): Long-running coding agents that need a million-token context. 1M context; $1.12 input, $0.14 cached input, $3.52 output per 1M tokens. Markdown: https://boundless.network/models/glm-5.3.md
- [GPT-5.6 Sol](https://inference.boundless.network/models/gpt-5.6-sol) (OpenAI): GPT-5.6 flagship for CLI-heavy coding and long-horizon problem solving. 262K context; $2.00 input, $0.20 cached input, $10.00 output per 1M tokens.
- [GPT-6 Sol](https://inference.boundless.network/models/gpt-6-sol) (OpenAI): Harder coding problems and agents that need to work for longer. 1M context; $2.00 input, $0.20 cached input, $10.00 output per 1M tokens.
- [Kimi K3](https://boundless.network/models/kimi-k3) (Moonshot AI): Complex coding and reasoning across long-running tasks. 1M context; $2.30 input, $0.23 cached input, $11.40 output per 1M tokens. Markdown: https://boundless.network/models/kimi-k3.md
- [Claude Opus 5.5](https://inference.boundless.network/models/claude-opus-5.5) (Anthropic): Long coding or research tasks that benefit from adaptive thinking. 1M context; $4.00 input, $0.20 cached input, $20.00 output per 1M tokens.
- [Claude Opus 5](https://inference.boundless.network/models/claude-opus-5) (Anthropic): Anthropic Opus 5 for demanding coding, review, and long-horizon agents. 262K context; $5.00 input, $0.50 cached input, $25.00 output per 1M tokens.
- [Claude Fable 5.1](https://inference.boundless.network/models/claude-fable-5.1) (Anthropic): Coding agents, front-end work, and research-heavy tasks. 262K context; $10.00 input, $0.25 cached input, $50.00 output per 1M tokens.
- [Laya](https://inference.boundless.network/models/convaiinnovations/laya) (Convaiinnovations/laya): Routing, classification, and guardrail decisions with calibrated probabilities. 8K context; — input, — cached input, — output per 1M tokens.

## Contact

- Talk to an engineer: https://boundless.network/#get-access