Claude Code Pro Drops to $17/yr; Cortecs Repricing Hits 20+ Models

Plans Anthropic overhauled Claude Code tiers: Pro annual price cut $200→$17/yr, Free and Max limits reworked to per-5-hour-session usage, and Team (Standard seat)/Team (Premium seat) plans replace the old Team - Standard/Premium seat offerings (evt 43217).

Prices A broad models.dev repricing wave hit coding models today. Biggest mover: vercel/deepseek-v4-pro input jumped $0.435→$1.74/MTok (+300%) (evt 42953), while cortecs/deepseek-v4-pro also rose $1.553→$1.73 (+11.4%) (evt 42722). Cuts included digitalocean/mimo-v2.5-pro input $0.8→$0.4 (-50%) (evt 43015), openrouter/deepseek-v4-flash input $0.14→$0.0882 (-37%) (evt 42925), and cortecs/kimi-k2.7-code input $1.28→$0.75 (-41.4%) (evt 42703). Qwen saw diverging moves: qwen3.6-27b input up $0.289→$0.6 (+107.6%) (evt 42902) versus qwen3-235b-a22b-2507 input down $0.1495→$0.09 (-39.8%) (evt 42891). Claude variants on cortecs mostly softened: claude-opus4-6 input $5.98→$5.313 (-11.2%) (evt 42735), claude-4-6-sonnet input $3.59→$3.196 (-11%) (evt 42734), claude-sonnet-4 input $3.307→$2.898 (-12.4%) (evt 42700). Meanwhile hyper/llama-3.3-70b-instruct input rose $0.5126→$0.6066 (+18.3%) (evt 42990) and cortecs/llama-3.3-70b-instruct input rose $0.089→$0.129 (+44.9%) (evt 42712). GPT-5.4 on cortecs eased slightly, input $3→$2.898 (-3.4%) (evt 42697).

Usage goose overtook GitLawb in OpenRouter coding token share, 22,656,827,433 vs 20,450,098,381 tokens over 7 days (+10.8% margin) (evt 43231).