Token Ledger Digest – 2026-05-22

#ai #llm #api #news

Token Ledger Digest – 2026-05-22

Lead: Z.ai: GLM 5.1 introduced pricing after being free; prompt $0.98/1M tokens, completion $3.08/1M tokens. Users who relied on the free tier now face a noticeable cost increase.

Added

Qwen: Qwen3.7 Max – prompt $2.50/1M, completion $7.50/1M. Relevant for teams needing long‑context (1M tokens) generation.
xAI: Grok Build 0.1 – prompt $1.00/1M, completion $2.00/1M. Useful for developers experimenting with xAI’s newest model.

Removed

Arcee AI: Trinity Large Preview – prompt $0.15/1M, completion $0.45/1M. Affects anyone who had this model pinned in workflows; consider alternatives.

Price Changes

inclusionAI: Ling-2.6-1T – prompt ↓ $0.30→$0.075/1M (‑75 %); completion ↓ $2.50→$0.625/1M (‑75 %). Cost‑sensitive applications benefit from the steep drop.
Qwen: Qwen3.6 Flash – prompt ↓ $0.1875→$0.12375/1M; completion ↓ $1.125→$0.7425/1M. Developers using flash variants see lower per‑token spend.
Qwen: Qwen3.6 35B A3B – prompt ↑ $0.149→$0.150/1M (≈+0.7 %); completion unchanged $1.00/1M. Negligible impact; monitor if usage scales.
Qwen: Qwen3.6 27B – prompt ↓ $0.32→$0.30/1M; completion unchanged $3.20/1M. Slight savings for mid‑size model users.
Qwen: Qwen3.6 Plus – prompt ↓ $0.325→$0.1794/1M; completion ↓ $1.95→$1.07315/1M. Notable reduction for plus‑tier workloads.
StepFun: Step 3.5 Flash – prompt ↓ $0.10→$0.09/1M; completion unchanged $0.30/1M. Minor edge‑case saving.