DEV Community

4663437Mehdi
4663437Mehdi

Posted on Originally published at 4663437mehdi.github.io

The Token Ledger Digest – 2026-09-08

The Token Ledger Digest – 2026-09-08

Lead change – biggest cost impact

MoonshotAI: Kimi Latest – Prompt price fell from $2.55 / 1M to $2.50 / 1M; completion price rose from $12.75 / 1M to $14.00 / 1M. Net increase of +$1.25 / 1M for output‑heavy workloads. Teams using Kimi Latest for long generations should re‑evaluate cost forecasts.

Notable price changes (sorted by absolute impact)

Model Prompt Δ ($/1M) Completion Δ ($/1M) Note
DeepSeek V3.1 –0.30 –0.70 Large cut; cheapest for chat‑style workloads.
DeepSeek V4 Pro 0423 +0.268 +0.535 Notable rise; consider alternatives for Pro tier.
Qwen3 14B +0.108 +0.670 Completion cost jumps sharply.
MoonshotAI Kimi K2.7 Code +0.050 +0.100 Moderate increase across both fields.
DeepSeek V4 Flash Latest +0.005 +0.070 Small upward tick; still among the cheapest flash models.
Z.ai GLM Flash Latest –0.004 –0.013 Minor reduction; retains low‑cost position.

Removed models (free tiers)

  • Z.ai: GLM 5.2 (free) – 256k context
  • MiniMax: MiniMax M3 (free) – 1,048,576 context
  • MiniMax: MiniMax M2.7 (free) – 196,608 context

These free offerings are no longer available; users must switch to paid plans or alternative providers.

Cheapest models today (per‑million)

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

Total models tracked: 428.

No other substantive changes were recorded.


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

Top comments (0)