BLOG

Notes from the inference layer.

Announcements, engineering deep-dives, and practical guides from the team building Blackbox.

8 min read

Nemotron on BLACKBOX: Open Weights, Encrypted Inference, and 420.2 tok/s

Blackbox is the orchestration layer for coding agents — unifying the best open- and closed-source models behind one secure, cost-efficient interface. Here is why Nemotron has become a cornerstone of the platform: frontier American open weights, 20–30× cheaper than closed-source, behind end-to-end encrypted inference, and 420.2 tok/s of concurrency-1 output.

READ ARTICLE
2 min read

TB v2.1 Blackbox — GPT-5.6 Sol + Opus 4.8 (90% pass@1)

A two-model Blackbox reaches 90.2% pass@1 on Terminal-Bench v2.1 (AA basis) — leading the Artificial Analysis leaderboard — by using a sandbox-executing critic to gate exactly one graded answer per task.

READ ARTICLE
2 min read

Benchmark Performance: Faster Inference, Reference-Level Model Quality

Serving a model through the BLACKBOX AI API gives higher throughput and lower latency with no observed loss in quality. Our inference stack speeds up token generation without retraining or changing model weights, and our benchmark runs show no observed regression against the published GLM 5.2, NVIDIA Ultra, and Kimi K2.7 references.

READ ARTICLE
2 min read

Artificial Analysis: BLACKBOX AI Is the #1 Fastest Nemotron 3 Ultra Provider

Artificial Analysis independently benchmarks every API provider serving NVIDIA Nemotron 3 Ultra. In the latest snapshot, BLACKBOX AI holds #1 output speed — 454.4 tokens per second, 47% ahead of the runner-up, at roughly a third of its price.

READ ARTICLE
2 min read

Docker OpenClaw E2EE

Point OpenClaw at a BLACKBOX AI end-to-end encrypted model from inside a Docker container. All requests are sealed with ECDH + AES-256-GCM before they leave your machine and only decrypted inside the GPU enclave.

READ ARTICLE
3 min read

Orchestrator–Executor: A Two-Agent Split That Beats a Solo Model on SWE Tasks

Split the agent in two — a strong orchestrator that plans and verifies, a cheaper executor that implements — and Terminal-Bench 2.0 scores jump from 58.4 to 69.7.

READ ARTICLE
12 min read

Blackbox Encrypted AI: Confidential LLMs at 22× Lower Cost

End-to-end encrypted LLM inference with hardware attestation, at $0.45 per million output tokens — 22× cheaper than GPT-4o, with confidentiality guarantees TLS alone can't provide.

READ ARTICLE
4 min read

Zero Data Retention: How Blackbox Enforces It By Default

Prompt and response content is not retained. Zero Data Retention is applied to every request by default, and where we can't meet it, we reject the request rather than serving it without protection — with usage metadata for billing retained transparently.

READ ARTICLE