Signal

Ideas shaping the agent economy.

Gemini 3.8 Flash listed at $0.75 per 1M input and $3.75 per 1M output, beside a corrected Gemini 3.6 Flash price

LatestSep 2026ModelsRouting

Gemini 3.8 Flash, and the Price We Had Wrong

Gemini 3.8 Flash is live on BlockRun at $0.75/$3.75 per 1M. Listing it made us re-read the price sheet, where Gemini 3.6 Flash had been billing twice what it should — and that stale number turned out to be hiding two fallbacks that were overcharging callers.

Read the article

A grid of dim dots with a scattered few lit bright, connected by a line to an image frame — 18B of 320B parameters active, reading a picture

Aug 2026Models

GLM-5.3 Flash: Multimodal at a Tenth the Price

Z.AI's GLM-5.3 Flash is live on BlockRun at $0.15/$0.50 per 1M tokens — a 1M-context model that reads images, for less than most flagships charge for text alone. What we probed before listing it, and the one number on Z.AI's pricing page we deliberately did not use.

Three video frames of increasing size on a horizontal axis, with dots of increasing size beneath each, illustrating resolution tiers as price tiers

Aug 2026Models

Grok Imagine Video 1.5: Resolution Is the Price

xAI's Grok Imagine Video 1.5 is live on BlockRun at $0.08/sec for 480p, $0.14/sec for 720p and $0.25/sec for 1080p — xAI's published rates, plus a flat tenth of a cent per clip. What the tier actually buys, and why the gate that enforces it has to be per-model.

Illustration of two identical video frames beside two price tags, one half the size of the other

Aug 2026Models

Seedance 2.0 Mini: Half the Rate, Same 720p

Seedance 2.0 Mini is live on BlockRun at $0.3996 for a 5-second 720p clip with synced audio — half what 2.0 Fast bills for the same output. Measured token rates, the duration ceiling, and the one thing that makes a cheaper tier worth defaulting to.

Terminal window with a curl command returning a chat completion, next to an open padlock

Aug 2026ModelsAgents

Free LLM API, No Key: Models You Can curl Right Now

Seven free, OpenAI-compatible LLM endpoints with no API key, no wallet, no signup — verified live today with a bare curl. Model list with context windows, quickstart in curl/Python/JavaScript, and the honest limits.

Smart LLM routing — 46 models, 14 dimensions, sub-1ms classification

Mar 2026RoutingModelsBenchmarks

LLM Router Benchmark: 46 Models, 8 Providers, Sub-1ms Routing

We benchmarked 46 LLM models from 8 providers through a production payment pipeline. Speed and intelligence are poorly correlated. Here's how we built a 14-dimension classifier that routes requests in under 1ms.

MiniMax M2.7 — the first self-evolving reasoning model, now live on BlockRun

Mar 2026Models

MiniMax M2.7 Is Live on BlockRun — The First Self-Evolving Reasoning Model

MiniMax M2.7 is now available on BlockRun. The first model that deeply participates in its own evolution — matching GPT-5.3-Codex on SWE-Pro at a fraction of the cost.