TL;DR — GPT-4o-mini costs $0.15 per million input tokens and $0.60 per million output tokens — OpenAI's budget workhorse tier. A typical chat request (1K in / 500 out) costs about $0.00045, meaning ~2,200 requests per dollar. It sits between the ultra-cheap open-weight tier (DeepSeek V4 Flash $0.14/$0.28) and mid-tier closed models, and remains one of the most-routed models for summarization, extraction, and light drafting.
Official pricing (as of August 2026)
| | Input /1M | Output /1M | |---|-----------|------------| | GPT-4o-mini | $0.15 | $0.60 | | Cached input | discounted (see OpenAI cache pricing) | — |
OpenAI discounts cached input tokens for repeated prompt prefixes; batch processing offers reduced rates for non-urgent workloads. Both matter at volume.
What this costs in practice
- Personal agent (50 requests/day, ~1K in / 500 out each): roughly $0.70/month
- Support bot (5K requests/day): roughly $67/month
- Bulk summarization (10M tokens in / 2M out): $2.70 per run
The pattern: 4o-mini is effectively free at personal scale and stays reasonable at product scale — its costs only become interesting at heavy pipeline volume, which is exactly when the open-weight tier below it starts to matter.
How it compares
| Model | Input /1M | Output /1M | Note | |-------|-----------|------------|------| | GPT-4o-mini | $0.15 | $0.60 | This page | | DeepSeek V4 Flash | $0.14 | $0.28 | Open-weight, ~half the output price | | Gemini 3 Flash | $0.075 | $0.30 | Cheapest input in class | | Claude Haiku 4.5 | $1.00 | $5.00 | Stronger, 7-8× the price (details) |
Where it fits in a routing strategy
In a routed setup, 4o-mini competes directly with Flash-class open models for the bulk tier — the offloading math usually favors whichever is cheapest per output token for your mix. ClawRouters includes it in the pool and routes to it when task difficulty and price line up.
Routing decision card
The exclusive layer: how a router actually treats this model — the analysis you will not find on a rate-card mirror site.
- Route here when: summarization, extraction, classification, light drafting where instruction-following matters more than reasoning depth
- Do not route here when: the same request qualifies for DeepSeek V4 Flash or Gemini 3 Flash at half the output price and quality parity — or when it needs mid-tier reasoning (escalate, do not retry)
- Head-to-head at this price point: DeepSeek V4 Flash ($0.14/$0.28) is the direct competitor; 4o-mini wins on ecosystem/format reliability, loses on output price
- Typical escalation pairing: 4o-mini first → Claude Haiku 4.5 on validation failure → mid-tier only if both miss
Prices from the provider's official rate card as of August 2026 — verify at the source before committing to volume, as rates change with model launches. Full market context: 2026 AI pricing guide · routing basics.