DEV Community

4663437Mehdi
4663437Mehdi

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

Token Ledger Digest – 2026-07-14

Token Ledger Digest – 2026-07-14

Added Models

  • Kwaipilot: KAT-Coder-Air V2.5 – New offering. Prompt $0.15/1M, Completion $0.60/1M. Who should care: Teams needing a large‑context (256k) coding assistant at mid‑range cost.
  • Kwaipilot: KAT-Coder-Pro V2.5 – New offering. Prompt $0.74/1M, Completion $2.96/1M. Who should care: Users prioritizing higher throughput and willing to pay premium for the same 256k context.

Removed Models (Free Tier)

  • LiquidAI: LFM2.5-1.2B-Thinking (free) – Removed.
  • LiquidAI: LFM2.5-1.2B-Instruct (free) – Removed.
  • OpenAI: gpt-oss-120b (free) – Removed. Who should care: Anyone relying on zero‑cost experimentation; must migrate to paid alternatives or other free models.

Price Changes

Model Change Old Prompt ($/1M) New Prompt ($/1M) Old Completion ($/1M) New Completion ($/1M) Who should care
Z.ai: GLM 5.2 ↓ Completion –$0.0828; slight Prompt ↓ –$0.0018 0.93 0.9282 3.00 2.9172 Cost‑sensitive users of GLM 5.2; savings on long completions.
Qwen: Qwen3.6 27B ↑ Prompt +$0.004; Completion unchanged 0.285 0.289 2.40 2.40 Developers using Qwen3.6 for prompt‑heavy workloads.
DeepSeek: V4 Flash ↑ Prompt +$0.013; ↑ Completion +$0.026 0.077 0.090 0.154 0.18 Teams balancing low‑cost flash inference; modest cost rise.
Google: Gemma 4 31B ↓ Prompt –$0.060; Completion unchanged 0.12 0.06 0.35 0.35 Users of Gemma 4 seeing halved prompt cost.
DeepSeek: V3.1 ↑ Prompt +$0.040; ↑ Completion +$0.160 0.21 0.25 0.79 0.95 Heavy‑usage applications; notable increase for completion‑intensive tasks.
OpenAI: gpt-oss-120b ↓ Prompt –$0.006; ↓ Completion –$0.030 0.036 0.030 0.18 0.15 Existing users benefit from lower overall cost.

Note: Per‑million‑token prices are derived by multiplying the per‑token rate by 1,000,000 and rounding to five decimal places for readability.

Cheapest models today (for reference: inclusionAI: Ling-2.6-flash ($0.01/1M prompt, $0.03/1M completion), IBM: Granite 4.0 Micro ($0.017/1M, $0.112/1M), Meta: Llama 3.1 8B Instruct ($0.02/1M, $0.03/1M).


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

Top comments (0)