The Confidential Development Environment (CDE) for the most sensitive code on Earth.
San Francisco, CA
Joined February 2026
- Tweets1.4K
- Following16
- Followers508
- Likes2.1K
Pinned Tweet
We built a cyber-capable AI model that, based on benchmarking and real-world testing, is perhaps the most capable offensive-security LLM we have seen
It scores 95% on Cybench, solving 38/40 tasks
But benchmarks only go so far, so we pointed it at real systems
Through authorized programs, we tested it against a Fortune-500 technology company, a major global telecom, and leading Web3 protocols.
The result: a 100% success rate - it found a viable path into every authorized target in our test set.
It adapts, reasons, and finds routes that aren’t obvious, even in systems that already passed audits and security reviews.
If this holds at scale, most current security assumptions break.
ORGN has removed models from Anthropic, OpenAI, SpaceX/xAI, Google, Meta, Microsoft and NVIDIA
Protecting the privacy and integrity of user data comes first.
Only open-source models and models we trust to maintain zero retention remain available.
Powerful AI should not make confidentiality optional.
Write code, debug issues and build your product, confidentially with ORGN CDE
orgn.com/download
Analyse transactions. Flag suspicious patterns. Keep customer data confidential.
Route sensitive tasks through ORGN Gate
gateway.orgn.com/
Confidential, zero-retention AI for scientists, available on ORGN Gate
gateway.orgn.com/
Route sensitive prompts by risk.
Frontier models with zero retention for average work
TEE models for sensitive workloads that require hardware-enforced isolation.
Both available on ORGN Gate
gateway.orgn.com
Before you give AI access to your apps, data, or code, ask:
Where is it processed?
Who can access it?
What is retained?
A privacy policy and pledge not to look are not always enough. Sometimes, for sensitive work, you need a system that can't look.
That, is ORGN.
We’ve made GLM-5.3 Flash on confidential computing hardware available through the ORGN Gateway.
Your prompts and sensitive data can now run encrypted in hardware-isolated environments, with attestation records you can verify.
For workloads where a privacy policy is not enough.
We’re sharing how GLM-5.3 helped build and optimize the inference infrastructure serving GLM-5.3-Flash.
The system went from its first successful run to production readiness in less than two weeks, with end-to-end throughput tripling relative to the initial baseline.
The key was dense feedback: local correctness tests, execution traces, microbenchmarks, and end-to-end measurements that enabled targeted hypothesis testing rather than reliance on aggregate performance metrics alone.
z.ai/blog/glm-built-its-infe…
Origin retweeted
We’re sharing how GLM-5.3 helped build and optimize the inference infrastructure serving GLM-5.3-Flash.
The system went from its first successful run to production readiness in less than two weeks, with end-to-end throughput tripling relative to the initial baseline.
The key was dense feedback: local correctness tests, execution traces, microbenchmarks, and end-to-end measurements that enabled targeted hypothesis testing rather than reliance on aggregate performance metrics alone.
z.ai/blog/glm-built-its-infe…
Three model types:
Abliterated
Private
Cyber-native
Because authorised security work shouldn't trigger arbitrary refusals
Because prompts and findings should stay confidential
Because defenders should have access to intelligence built explicitly for their line of work
And that's why we built ORGN on confidential computing
The point is not simply to promise privacy, but create an architecture where access is technically constrained
We effectively built a vault around customer workloads, then made sure to lock everyone, and even ourselves, out
Proud to be building this as part of the #NVIDIAInception program.
Enterprises should own their intelligence.
That means bringing AI to their sensitive data while protecting both business data and proprietary models.
NVIDIA’s @j_boitano and @VAST_Data Renen Hallak discuss how NVIDIA Confidential Computing makes this possible.
