H Company
on Orq.ai
Use H Company’s Holo model family through a single Orq.ai API. Route models such as Holo3‑122B‑A10B and Holo3‑35B‑A3B via Orq’s AI Router for agentic chat, reasoning, coding, and “computer‑use” workflows.
Capabilities:
Chat
Reasoning
Vision
Models Supported:
Holo3 122B A10B
Holo3.1 35B A3B
Provider HQ:
H Company, headquartered in Paris, France, with activity across Europe and beyond.
Access H Company through Orq.ai’s AI Router
H Company builds Holo, a family of mixture‑of‑experts foundation models optimized for agents that can operate desktops, browsers, and other applications, with variants targeting both large‑scale reasoning and efficient inference. These models cover high‑end reasoning, GUI/desktop agents, and cost‑efficient task automation.
H Company models available on Orq.ai
Model | Type | Context / scope | Best for | Pricing tier |
|---|---|---|---|---|
Holo3‑122B‑A10B | Chat / reasoning / computer‑use | Around 64K tokens (check H / Orq docs for current limits) | Complex reasoning, OSWorld‑style computer‑use benchmarks, and high‑stakes agent tasks where you want H Company’s flagship MoE model | Premium – priced around 0.40 USD per 1M input tokens and 3.00 USD per 1M output tokens via H Company’s API |
Holo3‑35B‑A3B | Chat / reasoning / agents | Around 262K tokens (verify in H / Orq docs) | Everyday agentic workloads, desktop automation, and product features that balance quality with cost and latency | Mid‑tier – priced around 0.25 USD per 1M input tokens and 1.80 USD per 1M output tokens via H Company’s API |
Holo‑1 / Holotron‑12B (where integrated) | Chat / vision / computer‑use | Multimodal context (GUI/screens; see H docs) | Lighter‑weight, multimodal “computer‑use” agents and browser automation where smaller models are sufficient | Cost‑efficient – public pricing focuses on larger Holo3 MoE models; lighter models are generally cheaper per 1M tokens when available via H’s platform |
Why use H Company through Orq.ai?
Capability | Provider / models | Direct | Through Orq.ai |
|---|---|---|---|
Chat / agents | Holo3‑122B‑A10B, Holo3‑35B‑A3B, Holo‑1 | Call Holo3 and related models directly via H Company’s API for chat, reasoning, and agentic control of desktops and browsers | Use H Company models through Orq.ai’s OpenAI‑compatible endpoint, adding routing, tracing, evals, budgets, and governance controls around each request |
Code | Holo3 models tuned or suitable for tooling/automation | Use H Company directly for agent workflows that include code manipulation, test automation (Tester H), and workflow orchestration (Runner H) | Route these workloads through Orq.ai, compare H Company Holo models against other providers, and monitor cost, latency, and quality from one control layer |
Embeddings / tools | Embedding providers plus H’s agents (Runner H, Surfer H, Tester H) | H Company focuses on agent and “computer‑use” models; embeddings and classic vector search are typically handled by other stacks | Use Orq.ai to route embeddings to dedicated providers while keeping Holo3 for reasoning, agent steps, and GUI automation |
This gives teams a practical way to use H Company where it performs best while centralising routing, observability, evals, and cost controls across the wider model stack.
Pricing
Model rates
H Company model pricing may differ depending on whether you:
connect your own H Company API key (BYOK), or
use H Company models billed through Orq.ai where available.
Key public figures include:
Holo3‑35B‑A3B: about 0.25 USD per 1M input tokens and 1.80 USD per 1M output tokens.
Holo3‑122B‑A10B: about 0.40 USD per 1M input tokens and 3.00 USD per 1M output tokens.
H’s agent products (for example Surfer H) are sometimes priced per task (for example around 0.13 USD per task at launch for certain benchmarks), but Orq.ai focuses on per‑token LLM usage when routing Holo models.
Check the Orq.ai pricing page and your workspace’s provider configuration for current per‑model rates, quotas, and plan details.
Compatible frameworks and tools
Orq.ai exposes H Company via:
a dedicated H Company provider configuration in the AI Router, and
an OpenAI‑compatible API layer for applications that expect that interface.
That means:
Popular AI frameworks (for example, OpenAI‑compatible clients and orchestration libraries) can talk to Holo3 via Orq’s router once the H Company provider is configured.
Code assistants and IDE tools that support MCP or OpenAI‑compatible APIs (Cursor, VS Code, Claude Desktop, Warp, Zed, and similar tools you’ve documented) can route through Orq.ai to H Company, depending on model and integration configuration.
Check the Orq.ai integration docs for the latest supported frameworks and tools for H Company.
FAQs
Do I need a separate H Company account to use H Company through Orq.ai?
You can either connect your own H Company API key into Orq.ai (from the H Company portal) or, where available, use H Company models billed via Orq.ai; the exact options depend on your Orq plan, region, and how H Company is configured in your workspace. In both cases, Orq.ai gives you one place to manage routing, observability, and cost controls around that H Company usage.docs.orq+3
Can I route only some workflows to H Company and others to different providers?
Yes. You define routes per workflow in Orq.ai and decide which ones should use Holo3 vs other models, so you can reserve H Company for desktop/browser agents and specific reasoning tasks while sending other workloads to different providers.orq+1
Does using H Company through Orq.ai add latency?
Orq.ai is designed as a lightweight router layer, so the added overhead is small compared to the model’s own latency. You can use routing policies, caching, and provider selection to keep end‑to‑end performance within your targets while leveraging H Company’s agent‑optimized models
Alternatives to
H Company
Anthropic
Use Claude Opus 4.7, Sonnet 4.6, Haiku 4.5, and other supported Claude models through one API.
Chat
Code
Reasoning
Vision
Models:
claude-opus-5
claude-sonnet-5
claude-fable-5
Open AI
Use OpenAI's foundation models through a single Orq.ai API. Route models such as GPT-4.1, GPT-4.1-mini, o3-mini, and GPT-4o-class models via Orq's AI Router for chat, reasoning, coding, and multimodal workloads.
Chat
Code
Image Generation
Reasoning
Speech
Vision
Models:
gpt-5.5 (EU)
gpt-5.6-luna (EU)
gpt-5.6-sol (EU)
Google AI
Use Google’s Gemini models through a single Orq.ai API. Route models such as Gemini 3.1 Pro, Gemini 2.5 Flash, and Gemini 2.0 Flash‑Lite via Orq’s AI Router for chat, reasoning, coding, and multimodal workloads.
Chat
Code
Embeddings
Image Generation
Reasoning
Speech
Vision
Models:
gemini-3.5-flash-lite (Gemini API)
gemini-3.6-flash (Gemini API)
Gemini 3 Pro Image
AWS
Use AWS Bedrock’s foundation models through a single Orq.ai API. Route models such as Amazon Nova, Amazon Titan Text, and compatible third‑party models exposed via Bedrock through Orq’s AI Router for chat, reasoning, coding, and multimodal workloads.
Chat
Code
Embeddings
Reasoning
Vision
Models:
eu.anthropic.claude-opus-5
global.anthropic.claude-opus-5
us.anthropic.claude-opus-5
