Cheaper Inference / Pricing history

Find the cheapest time to run each model.

Compare observed market prices with direct provider rates, then identify the hours that have historically delivered the strongest savings.

1d 7d 30d

Model to analyze

Time zone

aion-3.0

High confidence

This model averaged 43.6% below comparable direct provider prices across the selected range. Its strongest observed window was 3:00 AM–11:00 AM in UTC.

Last 30 days

Average saved

43.6%

Last 30 days

Cheapest time

3:00 AM–11:00 AM

45.0% average savings

Hourly coverage

24/24

281 total observations

Signal quality

High

Ready for planning

Typical unit savings$1.31 / 1M input tokens$2.61 / 1M output tokens

Hover a bar for savings per token and observation counts.

Source buckets are UTC; labels are converted to UTC.

Min savings

Showing 99 of 99 models

Confidence shows how much pricing history supports each model's best-time recommendation—not model quality.


deepseek-v4.1-flashDeepSeek	49.0%	6:00 AM	24/24	716	High confidence
glm-5Z.ai	46.1%	Collecting data	4/24	120	Early data
glm-4.6Z.ai	44.4%	1:00 AM–3:00 AM	24/24	712	High confidence
glm-4.5Z.ai	44.1%	5:00 AM–6:00 AM	24/24	711	High confidence
deepseek-v4-flashDeepSeek	43.8%	1:00 PM	24/24	715	High confidence
aion-3.0Aion Labs	43.6%	3:00 AM–11:00 AM	24/24	281	High confidence
glm-4.7Z.ai	43.5%	3:00 AM–4:00 AM	24/24	714	High confidence
glm-4.5-airZhipu AI	42.6%	1:00 AM	24/24	711	High confidence
deepseek-v4-flash-0731DeepSeek	39.9%	7:00 AM	24/24	716	High confidence
aion-3.0-miniAion Labs	39.8%	9:00 AM–10:00 AM	24/24	281	High confidence
deepseek-v4-pro-0813DeepSeek	35.2%	8:00 AM	24/24	715	High confidence
claude-haiku-5.5Anthropic	35.0%	7:00 PM–6:00 AM	12/24	12	Growing confidence
gemini-3-flashGoogle	34.9%	4:00 PM–12:00 AM	24/24	281	High confidence
glm-4.7-flashZ.ai	34.8%	4:00 PM–12:00 AM	24/24	310	High confidence
gemma-3-27b-itGoogle	34.7%	6:00 PM–1:00 AM	24/24	310	High confidence
gemma-4-26b-a4b-itGoogle	34.5%	2:00 AM	24/24	310	High confidence
gemini-3.5-flash-liteGoogle	34.4%	2:00 PM–6:00 PM	24/24	281	High confidence
gemini-3.5-flashGoogle	34.3%	2:00 PM–3:00 AM	24/24	281	High confidence
claude-opus-4.8-fastAnthropic	33.8%	1:00 AM–12:00 PM	24/24	281	High confidence
gemma-3-4b-itGoogle	33.6%	5:00 AM–8:00 AM	24/24	310	High confidence
deepseek-v4-proDeepSeek	33.4%	6:00 AM	24/24	715	High confidence
gemma-3-12b-itGoogle	33.1%	5:00 AM–6:00 AM	24/24	310	High confidence
gemini-3.8-flashGoogle	32.9%	5:00 AM	24/24	310	High confidence
claude-sonnet-5.5Anthropic	32.6%	3:00 PM	24/24	393	High confidence
claude-sonnet-4.5Anthropic	32.4%	12:00 AM–4:00 AM	24/24	716	High confidence
claude-opus-4.5Anthropic	32.1%	12:00 PM	24/24	715	High confidence
claude-opus-4.6Anthropic	32.0%	2:00 AM	24/24	716	High confidence
claude-sonnet-5Anthropic	32.0%	6:00 AM	24/24	715	High confidence
claude-fable-5Anthropic	31.9%	6:00 AM	24/24	715	High confidence
claude-opus-5Anthropic	31.8%	6:00 AM	24/24	715	High confidence
claude-opus-4.7Anthropic	31.7%	4:00 AM–6:00 AM	24/24	716	High confidence
claude-opus-4.8Anthropic	31.7%	4:00 AM	24/24	716	High confidence
claude-fable-5.1Anthropic	31.6%	2:00 AM	24/24	716	High confidence
claude-sonnet-4.6Anthropic	31.6%	1:00 AM	24/24	716	High confidence
claude-haiku-4.5Anthropic	31.5%	6:00 AM	24/24	716	High confidence
claude-opus-5-fastAnthropic	31.4%	8:00 AM	24/24	716	High confidence
claude-opus-5.5Anthropic	31.3%	12:00 AM	24/24	586	High confidence
gemini-2.5-flashGoogle	30.9%	10:00 AM	24/24	716	High confidence
gemma-4-31b-itGoogle	30.5%	5:00 AM	24/24	588	High confidence
gemini-3.7-flashGoogle	30.0%	11:00 PM	24/24	715	High confidence
gemini-3.1-flash-liteGoogle	29.6%	12:00 PM	24/24	716	High confidence
gemini-3.6-flashGoogle	27.9%	12:00 AM	24/24	707	High confidence
gemini-3.1-proGoogle	17.6%	10:00 PM	24/24	711	High confidence
glm-5.1Z.ai	--	Collecting data	0/24	0	Early data
glm-5.2Z.ai	--	Collecting data	0/24	0	Early data
glm-5.3Z.ai	--	Collecting data	0/24	0	Early data
glm-5.3-flashZ.ai	--	Collecting data	0/24	0	Early data
gpt-4.1-nanoOpenAI	--	Collecting data	0/24	0	Early data
gpt-5-imageOpenAI	--	Collecting data	0/24	0	Early data
gpt-5-miniOpenAI	--	Collecting data	0/24	0	Early data
gpt-5-nanoOpenAI	--	Collecting data	0/24	0	Early data
gpt-5.2-codexOpenAI	--	Collecting data	0/24	0	Early data
gpt-5.4OpenAI	--	Collecting data	0/24	0	Early data
gpt-5.4-miniOpenAI	--	Collecting data	0/24	0	Early data
gpt-5.4-nanoOpenAI	--	Collecting data	0/24	0	Early data
gpt-5.5OpenAI	--	Collecting data	0/24	0	Early data
gpt-5.5-proOpenAI	--	Collecting data	0/24	0	Early data
gpt-5.6-lunaOpenAI	--	Collecting data	0/24	0	Early data
gpt-5.6-solOpenAI	--	Collecting data	0/24	0	Early data
gpt-5.6-terraOpenAI	--	Collecting data	0/24	0	Early data
gpt-6-astraOpenAI	--	Collecting data	0/24	0	Early data
gpt-6-lunaOpenAI	--	Collecting data	0/24	0	Early data
gpt-6-solOpenAI	--	Collecting data	0/24	0	Early data
gpt-6.1-solOpenAI	--	Collecting data	0/24	0	Early data
gpt-image-2OpenAI	--	Collecting data	0/24	0	Early data
gpt-oss-120bOpenAI	--	Collecting data	0/24	0	Early data
grok-4.3xAI	--	Collecting data	0/24	0	Early data
grok-4.5xAI	--	Collecting data	0/24	0	Early data
grok-4.6SpaceXAI	--	Collecting data	0/24	0	Early data
grok-4.7SpaceXAI	--	Collecting data	0/24	0	Early data
grok-imaginexAI	--	Collecting data	0/24	0	Early data
jev-1.13TypeSafe	--	Collecting data	0/24	0	Early data
kimi-k2-thinkingMoonshot	--	Collecting data	0/24	0	Early data
kimi-k2.5Moonshot	--	Collecting data	0/24	0	Early data
kimi-k3Moonshot	--	Collecting data	0/24	0	Early data
mimo-v2.6-flashXiaomi	--	Collecting data	0/24	0	Early data
mimo-v2.6-proXiaomi	--	Collecting data	0/24	0	Early data
minimax-m2MiniMax	--	Collecting data	0/24	0	Early data
minimax-m2.1MiniMax	--	Collecting data	0/24	0	Early data
minimax-m2.5MiniMax	--	Collecting data	0/24	0	Early data
minimax-m2.7MiniMax	--	Collecting data	0/24	0	Early data
minimax-m3MiniMax	--	Collecting data	0/24	0	Early data
muse-spark-1.2Meta	--	Collecting data	0/24	0	Early data
nano-bananaGoogle	--	Collecting data	0/24	0	Early data
nano-banana-2Google	--	Collecting data	0/24	0	Early data
nano-banana-proGoogle	--	Collecting data	0/24	0	Early data
qwen3-32bQwen	--	Collecting data	0/24	0	Early data
qwen3-coder-30b-a3b-instructQwen	--	Collecting data	0/24	0	Early data
qwen3-coder-nextQwen	--	Collecting data	0/24	0	Early data
qwen3-next-80b-a3b-instructAlibaba	--	Collecting data	0/24	0	Early data
qwen3-vl-235b-a22b-instructQwen	--	Collecting data	0/24	0	Early data
qwen3.5-35b-a3bAlibaba	--	Collecting data	0/24	0	Early data
qwen3.6-27bAlibaba	--	Collecting data	0/24	0	Early data
qwen3.6-35b-a3bAlibaba	--	Collecting data	0/24	0	Early data
qwen3.8-2.4t-a95bQwen	--	Collecting data	0/24	0	Early data
qwen3.8-27bQwen	--	Collecting data	0/24	0	Early data
qwen3.8-flashQwen	--	Collecting data	0/24	0	Early data
qwen3.8-maxAlibaba	--	Collecting data	0/24	0	Early data
seedance-2.0ByteDance	--	Collecting data	0/24	0	Early data

Find the cheapest time to run each model.

aion-3.0

When this model is cheapest

Compare every model

Use the same models at market-linked rates

Product

Developers

Company

Legal