Token Ledger Digest – 2026-06-16

Most impactful change

• xAI: Grok 4.20 Multi-Agent – Prompt price dropped from $2.00/1M to $1.25/1M (−$0.75); completion price fell from $6.00/1M to $2.50/1M (−$3.50). Developers running multi‑agent workflows see roughly a 60% reduction in prompt cost and a 58% cut in completion cost.

Price changes

• DeepSeek: DeepSeek V4 Flash – Prompt rose $0.09 → $0.098/1M (+$0.008); completion rose $0.18 → $0.196/1M (+$0.016). Minor uptick; affects latency‑sensitive calls.
• Tencent: Hy3 preview – Prompt increased $0.063 → $0.066/1M (+$0.003); completion jumped $0.21 → $0.26/1M (+$0.05). Notable completion cost rise for long‑form generation.
• Qwen: Qwen3.5 397B A17B – Prompt slipped $0.39 → $0.385/1M (−$0.005); completion climbed $2.34 → $2.45/1M (+$0.11). Mixed; completion up ~5%.
• Z.ai: GLM 4.5 Air – Prompt edged up $0.125 → $0.13/1M (+$0.005); completion unchanged at $0.85/1M. Small prompt adjustment.

Model removed

• DeepSeek: R1 Distill Qwen 32B – No longer offered (previous rates $0.29/1M prompt & completion). Users must shift to alternatives; review any dependent workloads.

Cheapest models today (reference)

• inclusionAI: Ling-2.6‑flash – $0.01/1M prompt, $0.03/1M completion
• IBM: Granite 4.0 Micro – $0.017/1M prompt, $0.112/1M completion
• Meta: Llama 3.1 8B Instruct – $0.02/1M prompt, $0.03/1M completion

Total models tracked: 336.

---

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