DEV Community

4663437Mehdi
4663437Mehdi

Posted on Originally published at 4663437mehdi.github.io

The Token Ledger Digest – 2026-08-19

#ai #api #llm #news

The Token Ledger Digest – 2026-08-19

  • Z.ai: GLM 5.2 (batch) – Prompt: $0.70 → $1.40/M; Completion: $2.20 → $4.40/M (both doubled). Who should care: batch users seeing a 2× cost increase.
  • Thinking Machines: Inkling (batch) – Prompt: $0.50 → $1.00/M; Completion: $2.03 → $4.05/M (both doubled). Who should care: batch workloads now cost twice as much.
  • MoonshotAI: Kimi K2.7 Code (batch) – Prompt: $0.48 → $0.95/M; Completion: $2.00 → $4.00/M (both doubled). Who should care: batch API consumers facing a 2× price jump.
  • NVIDIA: Nemotron 3 Ultra (batch) – Prompt: $0.30 → $0.60/M; Completion: $1.80 → $3.60/M (both doubled). Who should care: batch users experiencing doubled expenses.
  • MiniMax: MiniMax M3 (batch) – Prompt: $0.15 → $0.30/M; Completion: $0.60 → $1.20/M (both doubled). Who should care: batch customers seeing a 2× rise.
  • Z.ai: GLM 5.2 – Prompt: $0.50 → $0.97/M (+94%); Completion: $3.15 → $3.04/M (−3.6%). Who should care: prompt‑heavy applications will notice higher costs.
  • Qwen: Qwen3.6 27B – Prompt: $0.60 → $0.30/M (−50%); Completion: $3.60 → $2.00/M (−44%). Who should care: cost‑sensitive users benefit from lower pricing.
  • Google: Gemma 4 31B – Prompt: $0.10 → $0.09/M (−10%); Completion unchanged at $0.34/M. Who should care: minor prompt savings for Google Gemma users.
  • Qwen: Qwen3.5-122B-A10B – Prompt: $0.29 → $0.26/M (−10%); Completion: $2.40 → $2.08/M (−13%). Who should care: moderate savings for both prompt and completion.
  • MoonshotAI: Kimi K2.5 – Prompt: $0.57 → $0.45/M (−21%); Completion: $2.85 → $2.25/M (−21%). Who should care: users see roughly a fifth‑price reduction.
  • Qwen: Qwen3 Next 80B A3B Instruct – Prompt: $0.10 → $0.09/M (−10%); Completion unchanged at $1.10/M. Who should care: slight prompt discount.
  • DeepSeek V4 Flash Latest – Prompt: $0.079 → $0.077/M (−2.7%); Completion: $0.157 → $0.153/M (−2.7%). Who should care: tiny cost decrease for high‑volume users.
  • Added: Z.ai: GLM 5.3 – Prompt: $1.40/M; Completion: $4.40/M. Who should care: new model option with mid‑range pricing.
  • Added: LiquidAI: LFM2.5-2.6B (free) – Prompt: $0.00/M; Completion: $0.00/M. Who should care: zero‑cost entry point for experimentation.

Originally published at The Token Ledger. Subscribe for the daily digest.

Top comments (0)