# DeAI News

> AI you control — open models, private inference, and the networks that run them. Independent daily coverage plus verification-graded provider trackers.

This file helps AI search engines and assistants understand and cite DeAI News. A clean Markdown version of any article is available by appending `.md` to its URL.

## Articles

- [Abliterated and Uncensored AI Models, Explained (2026)](https://www.deai.org/news/abliterated-uncensored-models-explained): Abliteration removes an LLM's refusal behavior by editing a single direction in its activation space — no retraining required. 'Uncensored' is the broader umbrella, covering abliterated checkpoints, fine-tunes on de-refused data, and steered base models.
- [The Best LLM APIs for AI Agents in 2026 (Tool Use, Filters, Priced)](https://www.deai.org/news/best-llm-api-for-ai-agents): Picking an LLM API for AI agents is about finishing multi-step tool-use loops, not topping chat benchmarks. This roundup compares seven providers — an aggregator, hosted open-weight platforms, and a decentralized inference marketplace — on function-calling reliability, filter.
- [The 10 Best Open-Source LLM API Providers in 2026 (Full Comparison)](https://www.deai.org/news/best-open-source-llm-api): There is no single best open-source LLM API — there are ten credible providers that split into aggregators, serverless generalists, speed specialists, and one decentralized inference marketplace.
- [The 8 Best Private AI APIs in 2026 (Retention Policies Compared)](https://www.deai.org/news/best-private-ai-api): All eight providers in this roundup state they do not train on your API prompts by default; they differ on how long they retain data and how you verify that. Rankings reflect documented policy terms as of 2026-08-20 — self-reported claims, not independent audits.
- [The 7 Best Uncensored AI APIs in 2026 (Refusal Policies Compared)](https://www.deai.org/news/best-uncensored-ai-api): Seven APIs serve open-weight models with minimal refusals in 2026, and they differ more by architecture than by any verified refusal rate. This roundup ranks them on documented content policies and uncensored model availability — no benchmark claims, because no public.
- [The Cheapest LLM APIs in 2026 — 14 Providers Priced per Million Tokens](https://www.deai.org/news/cheapest-llm-api): Fourteen LLM API providers compared on per-million-token pricing, with DeepSeek V4 Flash's $0.14/M input list price (as of 2026-08-20) as the verified low-end anchor. The real answer to 'cheapest' depends on your input/output mix, caching, and batch eligibility.
- [Chutes (Bittensor) vs Morpheus (2026): Decentralized AI Inference](https://www.deai.org/news/chutes-vs-morpheus): Chutes and Morpheus both serve open-weight models through OpenAI-compatible APIs, but they decentralize differently: Chutes is a subnet inside Bittensor's incentive network, while Morpheus is a standalone decentralized inference marketplace.
- [What Is Confidential AI Inference? TEEs and Who Offers It (2026)](https://www.deai.org/news/confidential-ai-inference-tee): Confidential AI inference uses hardware trusted execution environments (TEEs) to keep prompts encrypted while a GPU processes them, and remote attestation turns privacy from a policy promise into a checkable proof.
- [The 6 Decentralized AI Inference Networks That Actually Work in 2026](https://www.deai.org/news/decentralized-ai-inference-networks): Six decentralized AI inference networks are usable by builders in 2026: Chutes and Targon on Bittensor, Phala, Akash, Morpheus, and newer entrant Darkbloom. Most expose OpenAI-compatible endpoints, so evaluating one is usually a base-URL swap.
- [DeepSeek API: Official vs Third-Party Hosts (2026) — Privacy & Price](https://www.deai.org/news/deepseek-api-official-vs-third-party): DeepSeek's open weights and its official API are two different products: the weights can run anywhere, while the API sends your prompts to DeepSeek-operated servers under PRC jurisdiction.
- [Does Anthropic Train on Your Data? What the Policy Actually Says](https://www.deai.org/news/does-anthropic-train-on-your-data): Anthropic's answer to 'do you train on my data' splits in two: consumer Claude.ai conversations may be used if you've allowed it, while API and business-tier traffic is excluded by default per Anthropic's commercial terms.
- [Does DeepSeek Store Your Data? Jurisdiction and the Policy (2026)](https://www.deai.org/news/does-deepseek-store-your-data): DeepSeek's own privacy policy says the official app and website store user data — including prompts and uploaded files — on servers in the People's Republic of China, with open-ended retention.
- [Does Google Gemini Train on Your Data? What the Policy Says (2026)](https://www.deai.org/news/does-google-gemini-train-on-your-data): Google's answer depends on which Gemini you mean. Consumer Gemini apps may use your chats to improve Google products, with human review possible; the paid Gemini API, Vertex AI, and Workspace tiers carry terms saying your data is not used for training.
- [Does OpenAI Train on Your Data? What the Policy Actually Says (2026)](https://www.deai.org/news/does-openai-train-on-your-data): OpenAI's answer depends on the product: its policy excludes API and business-tier data from training, while consumer ChatGPT chats may be used unless you flip one Data Controls toggle. These are policy commitments, not independent audits — quoted and dated here as of 2026-08-20.
- [Does OpenRouter Log Your Prompts? The Router Nuance (2026)](https://www.deai.org/news/does-openrouter-log-your-prompts): Yes, OpenRouter processes and can retain your prompts — and so can every upstream provider it routes to. The useful question isn't whether one company logs, but how many copies of your prompt exist across the route. Both layers publish policies; neither policy is an audit.
- [Does xAI's Grok Train on Your Data? What the Policy Says (2026)](https://www.deai.org/news/does-xai-grok-train-on-your-data): Under xAI's consumer privacy policy, Grok can train on your conversations and X data by default — opting out is a single settings toggle, and API terms are a separate contract. The policy has churned repeatedly since launch, so every claim here is dated.
- [GPT-5.5 vs Open Models in 2026: Can DeepSeek V4, Kimi K3 Replace It?](https://www.deai.org/news/gpt-5-5-vs-open-models): As of August 2026, open-weight models like DeepSeek V4 and Kimi K3 can replace GPT-5.5 for most production workloads, typically at a fraction of frontier-API pricing. GPT-5.5's remaining moat is the hardest reasoning, long-horizon agentic, and multimodal work.
- [Leaving the Anthropic API: Open-Model Equivalents & the Switch (2026)](https://www.deai.org/news/migrate-from-anthropic): Migrating off the Anthropic API is mostly a base-URL swap plus a model shortlist: Qwen3, DeepSeek, and Llama cover most Claude workloads. Hosted zero-retention promises are policy statements; self-hosting is the only verifiable path.
- [Migrate Off the OpenAI API in an Afternoon (2026 — Code Included)](https://www.deai.org/news/migrate-from-openai): Migrating off the OpenAI API is a three-line code change: new base URL, new key, new model name — your existing SDK keeps working. The safe rollout is a 5% traffic canary for two hours with one-flag rollback, so the commitment is zero until the data says otherwise.
- [From Ollama to a Private Endpoint: Keep Privacy, Drop Ops (2026)](https://www.deai.org/news/ollama-to-private-endpoint): Ollama is excellent for local development, but production self-hosting means owning drivers, uptime, and a port that should never face the public internet.
- [Open-Weight vs Open-Source AI Models: The Difference That Bites (2026)](https://www.deai.org/news/open-weight-vs-open-source): Open-weight models hand you the files; open-source models hand you the freedoms. Most of 2026's flagship releases are the former wearing the latter's clothes, and the conditions only trigger at scale, by geography, or on redistribution. Read the license file before it reads you.
- [What Is an OpenAI-Compatible API? Why It Kills Vendor Lock-In (2026)](https://www.deai.org/news/openai-compatible-api): An OpenAI-compatible API is any inference endpoint that implements OpenAI's request and response format, so existing OpenAI SDK code runs against it after a one-line base_url change.
- [Top 9 OpenRouter Alternatives for Open Models (2026 — Priced)](https://www.deai.org/news/openrouter-alternatives): OpenRouter is a useful aggregator, but most teams serving open-weight models in production can go direct. These 9 providers all expose OpenAI-compatible endpoints, and switching is usually a one-line base-URL change.
- [OpenRouter vs Morpheus for Open Models (2026): Side by Side](https://www.deai.org/news/openrouter-vs-morpheus): OpenRouter is a centralized gateway to a broad model catalog with one API key; Morpheus is a decentralized inference marketplace where independent providers serve open-weight models.
- [How to Run DeepSeek V4 Flash in 2026: The $0.14/M Workhorse](https://www.deai.org/news/run-deepseek-v4-flash-api): DeepSeek V4 Flash is the budget tier of DeepSeek's V4 open-weight family, listed at $0.14 per million input tokens and $0.28 per million output as of 2026-08-20.
- [How to Run DeepSeek V4 Pro via API in 2026 — Every Host Compared](https://www.deai.org/news/run-deepseek-v4-pro-api): DeepSeek V4 Pro is reachable through the official DeepSeek API, third-party OpenAI-compatible hosts, aggregators, decentralized marketplaces, or your own GPUs. DeepSeek self-reports 80.6% on SWE-bench Verified as of 2026-08-20 — a vendor claim until independent evals land.
- [How to Run GLM-5.3 and GLM-5.2 via API in 2026 (Hosts Compared)](https://www.deai.org/news/run-glm-5-api): GLM-5.3 and GLM-5.2 are MIT-licensed open-weight models you can call through Z.ai's official API, third-party hosts, or your own hardware. GLM-5.3 shipped on 2026-08-14, so third-party availability still lags the official endpoint.
- [How to Run gpt-oss-120b via API in 2026 (Apache 2.0)](https://www.deai.org/news/run-gpt-oss-120b): gpt-oss-120b is OpenAI's 117B-parameter open-weight mixture-of-experts model, runnable behind any OpenAI-compatible API or self-hosted on a single 80 GB GPU.
- [How to Run Hermes 4 and Uncensored Fine-Tunes via API in 2026](https://www.deai.org/news/run-hermes-4-uncensored-models): Hermes 4 and uncensored fine-tunes like Dolphin are open weights, so any OpenAI-compatible host — or your own vLLM server — can serve them behind the same two-line client change.
- [How to Run Kimi K2.5 via API in 2026 (Hosts Compared)](https://www.deai.org/news/run-kimi-k2-5-api): Kimi K2.5 is served by Moonshot AI's first-party API and a half-dozen third-party hosts, nearly all behind OpenAI-compatible endpoints — reference pricing is about $0.60 per million input tokens and $3.00 per million output tokens as of 2026-08-20.
- [How to Run Kimi K3 via API in 2026 (Price, Context, License Caveats)](https://www.deai.org/news/run-kimi-k3-api): Kimi K3 is Moonshot AI's ~2.8-trillion-parameter open-weight MoE (~104B active, as reported by Moonshot as of 2026-08-20), callable from any OpenAI-compatible endpoint.
- [How to Run Llama 4 (Maverick & Scout) via API in 2026 — License Traps](https://www.deai.org/news/run-llama-4-api): Llama 4 Maverick and Scout are served by most major inference providers behind OpenAI-compatible endpoints, so integration is a base-URL swap.
- [How to Run Qwen3.6 (27B & 35B-A3B) in 2026: Local, API, or Both](https://www.deai.org/news/run-qwen-3-6): Qwen3.6 comes in two self-hostable sizes: a 27B dense model that fits one high-VRAM GPU and a 35B-A3B mixture-of-experts. Run it locally with Ollama, LM Studio, or vLLM, or call it through any OpenAI-compatible provider — switching is a base-URL change.
- [How to Run Qwen3 Coder 480B via API in 2026 (Apache 2.0)](https://www.deai.org/news/run-qwen3-coder-api): Qwen3 Coder 480B-A35B is an Apache 2.0 mixture-of-experts coding model you can call from almost any OpenAI-compatible provider by swapping a base URL and API key.
- [Self-Hosting vs Inference APIs in 2026: The Real Cost Math](https://www.deai.org/news/self-hosting-vs-api-cost): Self-hosting beats API pricing only above a break-even token volume that depends on your utilization, not on sticker prices. Below that line, per-token APIs win; above it, a well-loaded GPU server usually does.
- [Switching From OpenRouter to a Direct Provider (2026): When and How](https://www.deai.org/news/switch-from-openrouter): Switching from OpenRouter to a direct provider is usually a base-URL and API-key swap, not a rewrite — the hard part is knowing when router economics stop making sense. The trigger is typically sustained volume on one or two models, plus a need for direct data terms.
- [Together AI vs Fireworks AI (2026): Speed, Price, and Privacy Compared](https://www.deai.org/news/together-vs-fireworks): Together AI and Fireworks AI were both founded in 2022 and sell the same core product: fast, OpenAI-compatible serving of open-weight models, billed per million tokens.
- [What Is Decentralized AI Inference? A Builder's Guide (2026)](https://www.deai.org/news/what-is-decentralized-ai-inference): Decentralized AI inference replaces a single cloud vendor with a protocol that routes requests across independent GPU operators serving open-weight models.
- [What Does Zero Data Retention Actually Mean in AI APIs? (2026)](https://www.deai.org/news/zero-data-retention-ai-api): Zero data retention means a provider never persists your prompts or completions after serving a response — but it is a policy statement, not a verified fact, and coverage varies by tier, exceptions, and enforcement.
- [7 Zero-Retention AI APIs for Sensitive Workloads (2026)](https://www.deai.org/news/zero-retention-ai-apis): Seven hosted LLM APIs — OpenAI, Anthropic, Azure OpenAI, Amazon Bedrock, Google Vertex AI, Together AI, and Fireworks AI — publish no-training or zero-retention policies aimed at sensitive workloads.

## Topics

abliterated-models, ai-agents, akash-network, alibaba, anthropic, apache-2-0, api-migration, aws-bedrock, azure-openai, benchmarks-testing, bittensor, canary-deployment, cerebras, chatgpt, china-jurisdiction, chutes, claude, coding-agents, comparison, consumer-ai, context-caching, darkbloom, data-privacy, data-residency, data-retention, data-training, decentralized-inference, decentralized-infrastructure, deepinfra, deepseek, deepseek-v4, deepseek-v4-flash, deepseek-v4-pro, dolphin, featherless, fireworks-ai, function-calling, gemma-4, glm-5-2, glm-5-3, google, google-gemini, google-vertex-ai, gpt-5-5, gpt-oss, gpu-compute, gpu-infrastructure, grok, groq, hermes-4, hugging-face, hyperbolic, industry-business, inference-api, inference-apis, inference-pricing, inference-privacy, inference-providers, inference-routing, intel-tdx, kimi-k2-5, kimi-k3, licensing, llama-4, llama-4-maverick, llama-4-scout, llm-api, llm-inference, llm-routers, meta, migration-guides, mistral-ai, mixture-of-experts, model-comparison, model-evaluation, model-training, moonshot-ai, morpheus, nebius, nous-research, novita-ai, oasis-protocol, ollama, open-webui, open-weight-models, open-weights, open-weights-releases, openai, openai-api, openai-compatible-api, openrouter, opt-out, phala-network, privacy, privacy-policy, privacy-security, private-inference, provider-policy-trust, qwen, qwen3, qwen3-6, qwen3-coder, refusal-index, remote-attestation, sambanova, self-hosting, self-hosting-hardware, targon, tinfoil, together-ai, trusted-execution-environments, uncensored-models, vendor-lock-in, venice, venice-ai, vertex-ai, vllm, xai, z-ai, zero-data-retention, zero-retention

## Provider Tracker

https://www.deai.org/tracker — provider-level tracker of decentralized, private and open inference providers, verification-graded (claim vs. independently-checked fact). Machine access: https://www.deai.org/api/v1/providers (JSON), https://www.deai.org/api/v1/changes (append-only change feed).

## Open-Model Tracker

https://www.deai.org/models — open-weight model tracker: license, sizes, context window, benchmarks and price per million tokens across providers.

## Inference Price Index

https://www.deai.org/prices — weekly, dated snapshot of provider pricing per model, per million tokens.

## Refusal Index

https://www.deai.org/refusals — monthly standardized refusal testing across inference providers, published with methodology.

## Trust Tracker

https://www.deai.org/trust — provider data-retention, training-on-prompts, logging and jurisdiction comparison.

## Disclosure

DeAI: independent publication on open-weight AI models, private and decentralized inference. Maintains a provider tracker, a weekly open-model price index, and a monthly refusal index across inference providers — testing providers on price, privacy and refusal behavior and publishing the numbers, not the vibes. Powered by Morpheus (a decentralized inference marketplace, covered on the same terms as every other provider) and StrandCMS (the open-source, agent-first framework the site is built on).