DEV Community

4663437Mehdi
4663437Mehdi

Posted on • Originally published at 4663437mehdi.github.io

The Token Ledger Digest – 2026-08-17

#ai #api #llm #news

The Token Ledger Digest – 2026-08-17

Top Cost Impact

  • DeepSeek: DeepSeek V4 Pro 0813 – Prompt price ↑ from $0.435/M to $1.32/M (+$0.885/M); completion price ↑ from $0.87/M to $3.96/M (+$3.09/M). Who should care: Anyone using this model will see roughly a 4× increase in per‑million‑token cost.

Other Price Changes

  • Z.ai: GLM 5.2 – Prompt ↑ $0.308/M → $1.19/M (+$0.882/M); completion ↑ $0.968/M → $3.74/M (+$2.772/M).
  • NVIDIA: Nemotron 3.5 Lightning – Prompt ↓ $0.10/M → $0.08/M (−$0.02/M); completion ↓ $0.25/M → $0.20/M (−$0.05/M).
  • DeepSeek V4 Flash Latest – Prompt ↓ $0.067/M → $0.060/M (−$0.0068/M); completion ↓ $0.134/M → $0.121/M (−$0.0135/M).
  • MoonshotAI Kimi Latest – Prompt ↓ $2.80/M → $2.60/M (−$0.20/M); completion ↓ $14.0/M → $13.0/M (−$1.00/M).
  • Qwen: Qwen3.6 35B A3B – Prompt ↓ $0.15/M → $0.14/M (−$0.01/M); completion unchanged at $1.00/M.
  • Qwen: Qwen3.6 27B – Prompt ↓ $0.30/M → $0.289/M (−$0.011/M); completion ↑ $2.00/M → $2.40/M (+$0.40/M).
  • DeepSeek: DeepSeek V4 Pro 0423 – Prompt ↑ $1.168/M → $1.32/M (+$0.152/M); completion ↑ $2.336/M → $3.96/M (+$1.624/M).
  • DeepSeek: DeepSeek V4 Flash 0423 – Prompt ↑ $0.061/M → $0.083/M (+$0.021/M); completion ↑ $0.123/M → $0.165/M (+$0.042/M).
  • MoonshotAI: Kimi K2.6 – Prompt ↑ $0.542/M → $0.95/M (+$0.409/M); completion ↑ $2.28/M → $4.00/M (+$1.72/M).
  • Google: Gemma 4 26B A4B – Prompt ↓ $0.12/M → $0.07/M (−$0.05/M); completion ↓ $0.40/M → $0.34/M (−$0.06/M).
  • Qwen: Qwen3 VL 235B A22B Instruct – Prompt ↓ $0.26/M → $0.21/M (−$0.05/M); completion ↑ $1.04/M → $1.90/M (+$0.86/M).
  • Qwen: Qwen2.5 VL 72B Instruct – Prompt ↑ $0.25/M → $0.80/M (+$0.55/M); completion ↑ $0.75/M → $1.00/M (+$0.25/M).

Added Model

  • Z.ai: GLM 5.2 (free) – Prompt $0/M, completion $0/M, context 128k tokens. Who should care: Teams needing a no‑cost, large‑context model for prototyping or low‑budget workloads.

Cheapest Models Today (per‑million‑token)

  1. inclusionAI: Ling‑2.6‑flash – Prompt $0.01/M, Completion $0.03/M
  2. IBM: Granite 4.0 Micro – Prompt $0.017/M, Completion $0.112/M
  3. Mistral: Mistral Nemo – Prompt $0.019/M, Completion $0.030/M

Total models tracked: 414.

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

Top comments (0)