The Token Ledger – 2026-07-02
Most cost‑impacting change: Qwen Qwen3 30B A3B Thinking 2507 completion price rose from $0.40 to $1.56 per 1M tokens (+$1.16). Developers using this model for long‑form generation should reassess budget allocations.
Price changes
- DeepSeek V4 Flash – Prompt fell $0.098 → $0.089/1M (-$0.009); completion fell $0.196 → $0.18/1M (-$0.016). Who should care: Teams running high‑volume inference can save ~0.5 % on prompt and ~8 % on completion costs.
- MiniMax M2.1 – Prompt rose $0.29 → $0.30/1M (+$0.01); completion rose $0.95 → $1.20/1M (+$0.25). Who should care: Users of this model see a 26 % increase in completion expense.
- MiniMax M2 – Prompt unchanged at $0.255/1M; completion rose $1.00 → $1.02/1M (+$0.02). Who should care: Minor 2 % uplift for completion‑heavy workloads.
- Qwen Qwen3 30B A3B Thinking 2507 – Prompt rose $0.08 → $0.13/1M (+$0.05); completion rose $0.40 → $1.56/1M (+$1.16). Who should care: Completion cost nearly quadrupled; consider alternatives for token‑intensive tasks.
- Qwen Qwen3 8B – Prompt rose $0.05 → $0.117/1M (+$0.067); completion rose $0.40 → $0.455/1M (+$0.055). Who should care: Both prompt and completion costs up ~13‑14 %.
- DeepSeek Chat V3 0324 – Prompt rose $0.20 → $0.24/1M (+$0.04); completion rose $0.77 → $0.90/1M (+$0.13). Who should care: Completion expense up ~17 %; prompt up 20 %.
Cheapest models today (per 1M tokens)
- inclusionAI Ling‑2.6‑flash – Prompt $0.01, Completion $0.03
- IBM Granite 4.0 Micro – Prompt $0.017, Completion $0.112
- Meta Llama 3.1 8B Instruct – Prompt $0.02, Completion $0.03
No models were added or removed today.
Originally published at The Token Ledger. Subscribe for the daily digest.
Top comments (0)