OpenAI · GPT-4o · Small
GPT-4o mini pricing and limits
The low-cost GPT-4o variant, still a popular default for chat features and lightweight agents. Cheap per call, which is exactly why loops on it go unnoticed.
| Line | Tokens | Rate / 1M | Cost |
|---|---|---|---|
| Input, uncached | 1,500 | $0.15 | $0.00022 |
| Input, cache read | 1,500 | $0.075 | $0.00011 |
| Output | 400 | $0.60 | $0.00024 |
- Input / 1M
- $0.15
- Output / 1M
- $0.60
- Cached / 1M
- $0.075
- Cache write / 1M
- n/a
- Blended 3:1
- $0.26
- Context
- 128K
What drives GPT-4o mini cost
Per-token prices are low enough that teams rarely budget for it, yet a free-tier user hammering a chat feature can generate millions of tokens a month. The cost risk is volume and unmetered tenants, not price.
Cached input is billed at half the input rate. For many workloads GPT-5 nano or GPT-4.1 nano are now cheaper still; compare them in the price index before scaling a new feature on 4o-mini.
Cost of a typical call
Take a chat reply with 3,000 input tokens, 1,500 of them served from cache, and 400 output tokens. At GPT-4o mini list prices ($0.15 input, $0.075 cached input, $0.60 output per 1M) one call costs $0.00058, against $0.00069 with no cache hits. At 2,000,000 calls a month that is $1,155, or $1,380 uncached.
Blended at a 3:1 input to output ratio, GPT-4o mini costs $0.26 per 1M tokens. Routing the same call to the fallback in the policy below, GPT-4.1 nano, costs $0.00035, which is what a tenant pays you in COGS once they cross the hard cap and PALM SOUQ downgrades instead of blocking.
Cost comparison
GPT-4o mini against three alternatives
The same chat reply priced on GPT-5 nano, GPT-4.1 mini, Gemini 2.5 Flash-Lite. Move the sliders to match your traffic; models without a cache-read rate bill every input token in full.
Benchmark quality per task before switching. A cheaper model that needs more steps or longer outputs can cost more per finished task than its rate suggests, which is why PALM SOUQ reports cost per feature, not only per token.
GPT-4o mini this guide
$0.00058 per call
$1,444 / mo
baseline
GPT-5 nano
$0.00024 per call
$606 / mo
-58% vs GPT-4o mini
GPT-4.1 mini
$0.00139 per call
$3,475 / mo
+141% vs GPT-4o mini
Gemini 2.5 Flash-Lite
$0.00035 per call
$869 / mo
-40% vs GPT-4o mini
List prices from the PALM SOUQ LLM Price Index, computed per call as uncached input + cache reads + output. Reasoning tokens count as output on reasoning models.
Rate limits by usage tier
Organization and model level TPM and RPM by usage tier. High ceilings make it easy to burst, which helps throughput and hurts you when a loop starts.
Exact limits vary by account tier and change over time, so we do not reprint them here. Check the limits page for your own account in the OpenAI documentation at platform.openai.com/docs. PALM SOUQ enforces your own per-tenant limits in front of the provider's, so a single customer cannot consume the shared allowance that every other tenant depends on.
Circuit-breaker best practices
Token velocity is the breaker that matters here. A free-tier tenant burning tokens at many times its normal rate is either abuse or a loop, and either way the gateway should return HTTP 429 with X-PalmSouq-Reason: circuit_breaker_triggered.
Combine a per-tenant daily token quota with a soft warning webhook so your app can show the user a usage notice before a hard stop.
When a breaker trips, the gateway answers in under 50 ms with HTTP 429 and X-PalmSouq-Reason: circuit_breaker_triggered, before the request reaches OpenAI. Your agent framework sees a normal rate-limit response and stops, and the audit log records which heuristic fired and for which tenant.
# palmsouq.policy.yaml
policy: gpt-4o-mini-free
match:
model: gpt-4o-mini
tier: free # from X-PalmSouq-Tier
budgets:
monthly_usd: 5 # per tenant
soft_cap_pct: 80 # webhook to your app
on_hard_cap: fallback # or: block (HTTP 429)
fallback:
- gpt-4.1-nano
breakers:
tool_thrashing: { identical_calls: 3, window_steps: 8 }
token_velocity: { max_tokens_per_s: 50000, window_s: 10 }
semantic_echo: { repeats: 3 }
max_steps: 8
cache:
meter_reads: true- budgetsPer-tenant monthly dollar cap with an 80% soft-cap webhook, keyed by X-PalmSouq-Tenant.
- fallbackAt the hard cap, requests reroute to GPT-4.1 nano ($0.18 blended per 1M) instead of failing.
- breakersAll four loop heuristics on, evaluated in the request path with under 5 ms added latency.
Where it fits
Free tiers, autocomplete, intent routing and short chat answers. Keep it behind per-tenant quotas so a single heavy account cannot erase the margin of a whole plan.
Provider documentation: platform.openai.com/docs
GPT-4o mini pricing questions
How much does GPT-4o mini cost per million tokens?
At list price GPT-4o mini costs $0.15 per 1M input tokens and $0.60 per 1M output tokens, with cached input at $0.075 per 1M. Prices were last reviewed in the PALM SOUQ LLM Price Index and are for the standard tier and shortest context band.
Does GPT-4o mini support prompt caching?
Yes. Cached input is billed at $0.075 per 1M tokens against $0.15 for uncached input. PALM SOUQ records cached tokens per tenant from the provider's usage data, so cache savings show up in each customer's COGS.
What are the GPT-4o mini rate limits?
Organization and model level TPM and RPM by usage tier. High ceilings make it easy to burst, which helps throughput and hurts you when a loop starts. For your account's exact numbers, see the OpenAI documentation at platform.openai.com/docs.
How do I stop runaway GPT-4o mini agent loops?
Token velocity is the breaker that matters here. A free-tier tenant burning tokens at many times its normal rate is either abuse or a loop, and either way the gateway should return HTTP 429 with X-PalmSouq-Reason: circuit_breaker_triggered. The sample policy on this page enables all four PALM SOUQ heuristics with a max step cap of 8 and falls back to GPT-4.1 nano at the hard cap.
Is there a batch discount for GPT-4o mini?
50% off input and output through the Batch API (24-hour completion window). Batch traffic suits evaluations, backfills and nightly enrichment jobs that do not need a real-time answer.
Govern GPT-4o mini spend per customer.
Point your OpenAI client at gateway.palmsouqltd.com, pass X-PalmSouq-Tenant, and every GPT-4o mini token is attributed, capped and protected by circuit breakers.
14-day free production trial · 100k requests included · no card required