Models & pricing

Private by default. US-based model hosting and data storage.

22 open models behind one OpenAI-compatible endpoint (https://api.pgsgrove.com/v1). Prices are $ per 1M tokens. Pay per token from a prepaid balance with a $5 minimum, or take the coding plan, where a key is included and the same spend delivers up to 30% more usage. Cached input bills at the reduced cached rate automatically. Turbo ids route with priority at 1.5x the standard rate. See the quickstart to make your first call.

Model id	Modality	Context	Standard $/1M	Cached $/1M	Turbo $/1M
`mimo-v2.6-pro` MiMo V2.6 Pro · Xiaomi	Text + Image	1M	$0.56 in · $1.12 out	$0.005 cached in	n/a
The newest trillion-parameter flagship: deep reasoning, vision, and a million-token window, fresh off the September release. License: MIT.
`deepseek-v4.1-flash-1017` DeepSeek V4.1 Flash 1017 · DeepSeek	Text + Image	1M	$0.13 in · $0.37 out	$0.004 cached in	$0.195 in · $0.555 out
DeepSeek's October 17 update to V4.1 Flash: leaner thinking, quicker finishes, images and tools, a million-token window, at about half the price of V4.1 Flash. License: MIT.
`mimo-v2.6-flash` MiMo V2.6 Flash · Xiaomi	Text + Image	1M	$0.18 in · $0.36 out	$0.004 cached in	n/a
One of the quickest models on the roster: near-instant responses with a full million-token window, at a price that invites experimentation. License: MIT.
`ling-3.1-flash` Ling 3.1 Flash · inclusionAI	Text	256K	$0.38 in · $1.15 out	$0.08 cached in	n/a
Ant Group's newest Flash: a 560-billion-parameter hybrid reasoner with 25 billion active, built for agents, search, and real-world work, quick to first token. License: MIT.
`glm-5.2` GLM 5.2 · Z.ai	Text	1M	$0.65 in · $2 out	$0.12 cached in	$0.975 in · $3 out
Million-token context, strong tool use, priced well under the maker’s own rate. License: MIT.
`glm-5.3` GLM 5.3 · Z.ai	Text	1M	$1.6 in · $5.2 out	$0.32 cached in	n/a
Z.ai's newest full-size release: the flagship class pushed further on coding, security work, and long agentic runs, with the same million-token context. License: GLM-5.3 License.
`deepseek-v4.1-flash` DeepSeek V4.1 Flash · DeepSeek	Text + Image	1M	$0.26 in · $0.8 out	$0.01 cached in	$0.39 in · $1.2 out
DeepSeek's brand-new V4.1 Flash: the causal encoder-decoder release everyone is benchmarking, with a million-token window and a near-free cache read that rewards agentic loops. License: MIT.
`glm-5.3-flash` GLM 5.3 Flash · Z.ai	Text	1M	$0.19 in · $0.63 out	$0.04 cached in	n/a
Z.ai's newest Flash: the 5-series distilled into a light, million-token daily driver. License: MIT.
`kimi-k3` Kimi K3 · Moonshot AI	Text + Image	1M	$3.75 in · $15 out	$0.45 cached in	n/a
Moonshot's 2.8T flagship: a multimodal reasoning engine for long-horizon agentic work. The powerhouse option, priced like one. License: Kimi K3 License.
`deepseek-v4-flash-0731` DeepSeek V4 Flash 0731 · DeepSeek	Text	1M	$0.12 in · $0.23 out	$0.025 cached in	$0.18 in · $0.345 out
DeepSeek's 7/31 checkpoint of V4 Flash: a startling step up in benchmark scores at the same Flash price, with the same million-token context. License: MIT.
`deepseek-v4-pro-0813` DeepSeek V4 Pro 0813 · DeepSeek	Text	1M	$1.65 in · $3.3 out	$0.13 cached in	$2.475 in · $4.95 out
The official V4 Pro release: sharper agentic work and faster serving at the same price as the preview. License: MIT.
`kimi-k2.6` Kimi K2.6 · Moonshot AI	Text + Image	256K	$0.95 in · $4.4 out	$0.19 cached in	$1.425 in · $6.6 out
Proven Kimi generalist with strong tool use and image input. License: Modified MIT.
`minimax-m3` MiniMax M3 · MiniMax	Text + Image	512K	$0.38 in · $1.5 out	$0.075 cached in	n/a
A highly competent coding model from MiniMax. (Built with MiniMax M3.) License: MiniMax Community.
`deepseek-v4-pro` DeepSeek V4 Pro · DeepSeek	Text	1M	$1.65 in · $3.3 out	$0.13 cached in	$2.475 in · $4.95 out
Frontier reasoning and coding at a million-token window. License: MIT.
`qwen-3.8-27b` Qwen 3.8 27B · Qwen	Text	256K	$0.52 in · $3.9 out	$0.052 cached in	n/a
The compact sibling of the Qwen 3.8 flagship: 27B distilled from the same generation, fast and inexpensive for everyday coding and chat. License: Apache 2.0.
`deepseek-v4-flash` DeepSeek V4 Flash · DeepSeek	Text	1M	$0.12 in · $0.23 out	$0.025 cached in	$0.18 in · $0.345 out
Fast, low-cost DeepSeek with the same million-token context. License: MIT.
`nemotron-3-ultra` Nemotron 3 Ultra · NVIDIA	Text + Image	256K	$0.65 in · $2.8 out	$0.13 cached in	$0.975 in · $4.2 out
NVIDIA’s flagship reasoning model, with image input. License: OpenMDW 1.1.
`nemotron-3.5-super-vl` Nemotron 3.5 Super VL · NVIDIA	Text + Image	256K	$0.13 in · $0.5 out	$0.1 cached in	n/a
NVIDIA’s newest Super: reads images, screenshots and charts alongside text, thinks before it answers, and runs fast at a small-model price. License: OpenMDW 1.1.
`nemotron-3-super` Nemotron 3 Super · NVIDIA	Text	256K	$0.11 in · $0.5 out	n/a	$0.165 in · $0.75 out
Frontier-class reasoning quality at bargain-tier pricing. License: NVIDIA Open Model.
`gemma-4-31b` Gemma 4 31B · Google	Text + Image	256K	$0.25 in · $0.5 out	n/a	$0.375 in · $0.75 out
Efficient dense multimodal model with great price to capability. License: Gemma (Apache 2.0).
`kokoro-82m` Kokoro 82M · Hexgrad	Text	tts	$1.25 / 1M characters	n/a	$0 in · $0 out
Natural, fast text to speech from a tiny open model. License: Apache-2.0.
`embeddinggemma-300m` EmbeddingGemma 300M · Google	Text	embedding	$0.0025 in · $0 out	n/a	n/a
Compact 768-dimension embeddings for search and retrieval. License: Gemma Terms.