Blog
Technical notes from TopxAI: model introductions, endpoint walkthroughs and what changed in the catalogue.
- · GPT-6.1 Terra on TopxAI · Half the price of GPT-6.1 Sol, image input, a 1.05M-token context and reasoning you can switch off: prices, settings and a Responses example.
- · Moving to deepseek-flash-Abliterated on TopxAI · Flash moves to a private deployment with a new model ID, fixed $1.20 / $0.15 / $2.40 token prices, image inputs and a 262,144-token output limit.
- · Kimi K3.1 Pro replaces Kimi K3 on TopxAI · Move to kimi-k3.1-pro with unchanged token prices, a 1,048,576-token context and always-on reasoning. The old ID is retired.
- · What we fixed in the relay: Claude Code images, fairer billing, fewer failed requests · Claude Code image reads on non-Claude models, no charge for an image answer without an image, cache billing fixes and the requests that stopped failing.
- · Jev 1.14 on TopxAI: images, Box questions and what they cost · Jev 1.14 replaces 1.13 at the same price and can look at images. Our measurements of Box accuracy, image token counts and the real size limits.
- · Moving to Claude Sonnet 5.5 on TopxAI · Update the Sonnet model ID, keep the same input and output prices, and check thinking and tool settings when moving your client to Claude Sonnet 5.5.
- · Moving to GPT-6.1 Sol on TopxAI · The new model ID, lower cache-read prices, reasoning settings and a Responses API example for the GPT-6.1 Sol upgrade.
- · AI coding agents are moving the bottleneck to verification · What recent research says about agent productivity, why more output can slow delivery, and a practical evaluation worksheet for small engineering teams.
- · An AI API timeout is an unknown outcome, not a failed action · How to separate generation retries from tool execution, recover after an interrupted stream, and test duplicate effects before shipping an agent.
- · Why I paused an AI project: the cache bill missing from the architecture · Jevis had a detailed design for frontier-model planning and cheaper execution. Prompt-cache economics exposed the hypothesis we still had to test.
- · TypeSafe Jev: a System One model that returns judgments, not text · What Jev (jev-1.14.0) is, how its noul, choice, score and Box questions work, where it is weak, what it costs, and how to call it through TopxAI.
TopxAI · Models & pricing · Docs
Other languages: 简体中文