daily brief ·

Claude Code Pro Jumps to $200/yr as Nano-GPT Cuts GPT-5-Codex 87.5%

Plans Claude Code overhauled its pricing: Pro annual jumped $17→$200/yr, Free/Team-Standard/Team-Premium/Enterprise prices are no longer listed (previously $0/yr, $20/yr, $100/yr, $20/mo respectively), and the Education tier was removed (evt 41275).

Usage On OpenRouter's coding leaderboard, GDevelop overtook Portkey AI, 60,732,170,246 vs 54,131,432,044 tokens over 7 days (+12.2% margin) (evt 41288).

Prices DigitalOcean cut across its lineup: glm-5.2 -33.3% input/-50% output (evt 40825), deepseek-3.2 -41.2% (evt 40820), deepseek-v4-pro -37.5% (evt 40816), deepseek-4-flash -25% (evt 40811), llama-4-maverick -20% (evt 40812), gpt-oss-120b -21.4% (evt 40822), qwen3.5-397b-a17b -21.4% (evt 40818), nvidia-nemotron-3-super-120b -21.4% (evt 40824). Qwen2.5-vl-72b-instruct input fell 68.7% across OpenRouter, Kilo and models.dev (evts 40968, 40879, 40605), while mistral-small-3.2-24b-instruct rose 25% on the same three surfaces (evts 40946, 40855, 40601). Kimi-K2.6-TEE input eased 12.1% (evt 41021); mimo-v2.5-pro input rose 33.3% (evt 41071).

On nano-gpt, gpt-5-codex input crashed 87.5% to $1.25/MTok (evt 38868), claude-sonnet-latest fell 33.2% (evt 38835), and ring-2.6-1t dropped 70% (evt 38840). Meanwhile moonshotai's kimi-k2 family repriced sharply higher: kimi-k2-instruct variants +300% (evts 38846, 38848, 38851), kimi-latest +400% (evt 38845), kimi-k2-thinking +100% (evt 38843); grok-latest rose 60% (evt 38839) and mixtral-8x22b +122.2% (evt 38838).

events in this brief

On the tape

40 detected events · each links its primary source