# Blog — Hanzo AI

> AI infrastructure insights: agent architecture, Zen model releases, model routing guides, MCP tools, and enterprise AI deployment. From the team building the AI cloud.

Blog

# News& Guides.

Product launches, architecture deep dives, and how-to guides from the team building the AI workforce platform.

[Decision ModelsSeptember 30, 2026 ## What Is a Decision Model? Generative models reason in unconstrained paragraphs. Decision models decide in typed, calibrated state transitions. Here is why the next frontier of AI is not generation, but decisions.](https://hanzo.blog/blog/2026-09-30-what-is-a-decision-model/)

[KaiSeptember 28, 2026 ## Kai 1: Decisions, Not Completions Kai is Hanzo's decision model, served by the Hanzo API. Give it explicit state and a listed question, and it returns a typed, calibrated answer — a choice, a score, a yes/no, or a reasoned deferral.](https://hanzo.blog/blog/2026-09-28-introducing-kai/)

[Enterprise AISeptember 21, 2026 ## Top Pain Points for Investment Firms That Hanzo AI Solves Identify key operational challenges in private equity, venture capital, and asset management — and discover how Hanzo's sovereign AI cloud, frontier reasoning models, and agentic workflows streamline sourcing, diligence, and execution.](https://hanzo.blog/blog/2026-09-21-ai-pain-points-investment-firms/)

[architectureJuly 27, 2026 ## The router is the registry Hanzo Cloud has no checked-in OpenAPI file. The live route table is the source, projected three ways, and a bijection test fails the build if any projection drifts. 1,362 operations across 963 paths, generated per process at request time.](https://hanzo.blog/blog/2026-07-27-the-router-is-the-registry/)

[infrastructureJuly 17, 2026 ## Half a Million RPS on a GB10: The Network Was the Bottleneck, Not the Box We set out to see how fast our cloud data plane could go on an NVIDIA GB10. The answer taught us more about NIC receive queues than about our own code — and it more than doubled our throughput on the same wire.](https://hanzo.blog/blog/2026-07-17-half-million-rps-gb10-zip-zap/)

[ensoJuly 17, 2026 ## Introducing Enso — one API, routed to the best frontier intelligence Enso is Hanzo's learned router: no single frontier model leads every benchmark, so Enso routes each request to the model most likely to win it — and learns from your feedback as it goes. Microsecond routing on a CPU, a transparent per-request meter, and a 1% fee that pays for itself. Here is how it works, the measured results, and the cost-transparency angle a monolith cannot offer.](https://hanzo.blog/blog/introducing-enso/)

[routerJuly 7, 2026 ## Route every prompt to the right model — and cut AI cloud costs by up to 90% Hanzo Router places each request on the cheapest model that can still serve it — reuse a local model for free, load one that fits, or fall through to cloud. Here's the idea, what it feels like in production (qualitative), and the honest cost arithmetic (quantitative) — no fabricated benchmark.](https://hanzo.blog/blog/2026-07-07-hanzo-router-cut-ai-costs/)

[datastoreJune 26, 2026 ## Introducing Hanzo Datastore: a post-quantum-ready OLAP engine A ClickHouse-derived columnar warehouse, rebuilt around storage/compute separation on Hanzo S3, profile-guided and link-time optimized to roughly half the memory, with gRPC stripped out and a path to leaderless post-quantum coordination.](https://hanzo.blog/blog/2026-06-26-introducing-datastore/)

[datastoreJune 26, 2026 ## Killing Raft: from ZooKeeper consensus to Lux Quasar, natively How Hanzo Datastore replaces the NuRaft engine under ClickHouse Keeper with Lux Quasar — post-quantum, leaderless consensus linked directly into the C++ server — while keeping the ZooKeeper API that replication speaks.](https://hanzo.blog/blog/2026-06-26-killing-raft-quasar-native-coordination/)

[proof-of-aiJune 26, 2026 ## Proof of AI: An Open-Source Cloud That Pays the Developers It Runs On Hanzo Cloud is built on Proof of AI — deterministic inference any validator can re-verify on-chain — and on a commitment that 25% of what the cloud earns flows back to the open-source developers whose code it runs.](https://hanzo.blog/blog/proof-of-ai/)

[brainMay 12, 2026 ## Hanzo Brain: Local-First Knowledge Graph for Agents A single SQLite file your agent reads from every runtime. Zero LLM round-trips for graph ops, FTS5 plus vector plus typed-edge extraction at write time.](https://hanzo.blog/blog/2026-05-12-hanzo-brain/)

[zenApril 6, 2026 ## Zen: 1T+ Parameter MoE Architecture Hanzo's flagship LLM — a 1.04 trillion parameter Mixture of Experts model fused from top language models, with distributed training and on-chain model attestation.](https://hanzo.blog/blog/2026-04-06-zen-model-architecture/)

[formal-verificationMarch 26, 2026 ## Hanzo Ships Formal Proofs for AI Agent Infrastructure Machine-checked Lean 4 proofs for agent safety, MCP protocol, PoAI consensus, and platform operations](https://hanzo.blog/blog/2026-03-26-formal-proofs-ship/)

[aiFebruary 28, 2026 ## BitDelta: Serving 14 Zen Models from One GPU Cluster How BitDelta's 1-bit delta compression lets us serve 14 Zen model variants from shared GPU infrastructure — the math, the architecture, and the tradeoffs.](https://hanzo.blog/blog/2026-02-28-bitdelta-multi-tenant/)

[hanzoFebruary 28, 2026 ## From Commerce AI to Frontier Models: Hanzo's Decade of Infrastructure Twelve years of building AI infrastructure — from a real-time recommendation engine in 2014 to 94+ frontier models and a post-quantum blockchain in 2026. The technical story of how we got here.](https://hanzo.blog/blog/2026-02-28-decade-of-hanzo/)

[cloudFebruary 27, 2026 ## Enterprise Cloud Without Enterprise Pricing: 96 vCPUs at $3,999/Month Hanzo Network launches Business through Ultra enterprise cloud tiers — dedicated CPU machines from 8 to 96 vCPUs with 71–81% cost advantage over hyperscaler equivalents.](https://hanzo.blog/blog/2026-02-27-enterprise-cloud-tiers/)

[aiFebruary 27, 2026 ## The Complete AI Agent Stack: Models, Compute, and Tools in One Platform Hanzo AI introduces the first platform combining 100+ AI models, cloud compute, GPU access, and 260+ MCP tools under a single developer account.](https://hanzo.blog/blog/2026-02-27-full-stack-ai-agents/)

[gpuFebruary 27, 2026 ## H100 GPUs at $3.48/Hour — 72% Cheaper Than AWS Hanzo Bot opens on-demand H100 GPU access for AI agent developers at $3.48/hr — 72% less than AWS equivalent pricing.](https://hanzo.blog/blog/2026-02-27-gpu-access-h100/)

[botsFebruary 27, 2026 ## Run AI Agents for $5/Month: Introducing Hanzo Bot Hanzo Bot launches with the cheapest AI agent hosting in the industry — starting at $5/mo with free credit, integrated access to 100+ models, and H100 GPUs at $3.48/hr.](https://hanzo.blog/blog/2026-02-27-hanzo-bot-launch/)

[cloudFebruary 27, 2026 ## Hanzo Network: 10–54% Cheaper Than DigitalOcean, Lightsail, and Vultr Hanzo Network launches a developer cloud that undercuts DigitalOcean, AWS Lightsail, Vultr, and Linode by 10–54% — with zero egress fees, DDoS protection, and consistent global pricing included.](https://hanzo.blog/blog/2026-02-27-hanzo-network-cloud-launch/)

[pricingFebruary 27, 2026 ## AI Transcription 85% Cheaper Than OpenAI: Multimodal Pricing for Bot Developers Hanzo Bot delivers speech-to-text at $0.0009/minute — 85% cheaper than OpenAI Whisper — plus image generation at the industry's lowest price of $0.00013/step.](https://hanzo.blog/blog/2026-02-27-multimodal-pricing/)

[pricingFebruary 27, 2026 ## Hanzo Pricing Deep Dive: Every Model, Every Plan, Every Comparison The comprehensive breakdown of Hanzo pricing across AI models, cloud compute, GPU access, and tools — with head-to-head comparisons against every major competitor.](https://hanzo.blog/blog/2026-02-27-pricing-deep-dive/)

[aiFebruary 27, 2026 ## One API for Every AI Model: Introducing the Hanzo AI Gateway Hanzo AI launches the industry's first zero-markup multi-provider AI gateway — one API key for 100+ models from every major provider, plus 14 proprietary Zen models.](https://hanzo.blog/blog/2026-02-27-unified-ai-gateway/)

[aiFebruary 27, 2026 ## Zen4 Coder: 480B Parameters, 92% Fewer Active — The MoE Code Model Hanzo AI launches Zen4 Coder, a 480B-parameter Mixture of Experts code model that activates only 35B parameters per token — delivering frontier code intelligence at a fraction of the compute cost.](https://hanzo.blog/blog/2026-02-27-zen4-coder-launch/)

## Build with Hanzo

Deploy AI agents that code, market, and operate — starting free.

[Every post](https://hanzo.blog)[Read the Docs](https://docs.hanzo.ai)