DEV Community

4663437Mehdi
4663437Mehdi

Posted on Originally published at 4663437mehdi.github.io

Token Ledger Digest – 2026-09-13

#ai #llm #api #news

Token Ledger Digest – 2026-09-13

12 price changes, no additions or removals.

  • DeepSeek: DeepSeek V4 Pro 0423 – prompt rose from $0.79/1M to $1.60/1M; completion rose from $1.58/1M to $3.20/1M. Developers running high‑volume inference will see ≈+$2.43 per 1M tokens.
  • MoonshotAI: Kimi K3 – prompt rose from $2.30/1M to $2.65/1M; completion rose from $11.55/1M to $13.28/1M. Cost‑sensitive apps using this model face ≈+$2.08 per 1M tokens.
  • Z.ai: GLM 5.3 – prompt fell from $1.40/1M to $1.09/1M; completion fell from $4.40/1M to $3.43/1M. Users benefit ≈‑$1.28 per 1M tokens.
  • MoonshotAI: Kimi Latest – prompt fell from $2.13/1M to $2.10/1M; completion fell from $11.90/1M to $10.95/1M. Savings ≈‑$0.98 per 1M tokens for existing workloads.
  • Meta: Llama 3.1 70B Instruct – prompt fell from $0.72/1M to $0.40/1M; completion fell from $0.72/1M to $0.40/1M. Reduction ≈‑$0.64 per 1M tokens.
  • Meta: Muse Glimmer 30B – prompt rose from $0.30/1M to $0.35/1M; completion rose from $1.10/1M to $1.50/1M. Increase ≈+$0.45 per 1M tokens.
  • Z.ai: GLM 5.3 Flash – prompt rose from $0.08/1M to $0.15/1M; completion rose from $0.25/1M to $0.50/1M. Increase ≈+$0.33 per 1M tokens.
  • Z.ai: GLM Latest – prompt rose from $0.87/1M to $0.94/1M; completion fell from $3.36/1M to $3.17/1M. Net change ≈‑$0.13 per 1M tokens.
  • Google: Gemma 4 26B A4B – prompt rose from $0.04/1M to $0.09/1M; completion rose from $0.22/1M to $0.30/1M. Increase ≈+$0.13 per 1M tokens.
  • Qwen: Qwen3 30B A3B Instruct 2507 – prompt fell from $0.09/1M to $0.05/1M; completion fell from $0.30/1M to $0.19/1M. Savings ≈‑$0.15 per 1M tokens.
  • DeepSeek: DeepSeek V4 Flash Latest – prompt rose from $0.03/1M to $0.04/1M; completion rose from $0.07/1M to $0.11/1M. Increase ≈+$0.04 per 1M tokens.
  • DeepSeek: DeepSeek V4 Flash 0423 – prompt fell from $0.07/1M to $0.05/1M; completion fell from $0.13/1M to $0.10/1M. Reduction ≈‑$0.05 per 1M tokens.

Cheapest models today (per‑token prices):

  1. IBM: Granite 4.0 Micro – $0.017/1M prompt, $0.112/1M completion
  2. Mistral: Mistral Nemo – $0.019/1M prompt, $0.030/1M completion
  3. inclusionAI: Ling 3.0 Flash – $0.021/1M prompt, $0.063/1M completion

Developers should monitor the highlighted changes for budget impact, especially the sharp rise in DeepSeek V4 Pro and MoonshotAI Kimi K3.

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

Top comments (0)