OpenAI Models

Q: Which OpenAI model should I start with?

auto is free on input — the cheapest entry here that declares tool calling, and it carries a 1M context. Move up to o1-pro when answer quality matters more than cost, or to gpt-5.5-free for long-form reasoning.

Q: How is cached input billed?

The Cache read column is the rate for input tokens served from the prompt cache — for example gpt-5 bills cache hits at 10% of the input rate and gpt-5-chat-latest bills cache hits at 10% of the input rate. Cache write is the surcharge for putting a prompt into the cache in the first place, and only a few upstreams bill it separately. A dash in either column means the catalog carries no cache rate for that model, so plan on paying the full input rate.

134 modelsGeneral models free to startUp to 1.05M context

Usage

Last 29 days · 2026-08-10 to 2026-09-07

Tokens

1003B

Requests

74M

Models in use

90 of 134

Tokens per day, stacked by model

gpt-5.6-luna
gpt-5.6-sol
gpt-4o-mini
gpt-5.6-terra
omni-moderation-latest
gpt-5.5
gpt-5.4-mini
gpt-5.4
72 more models

Which models that traffic went to

Share of 1003B tokens. 10 models with traffic report no token counts and cannot be ranked here, including auto and gpt-4o-audio-preview — they are in the request view.

The two views disagree on purpose: a model can take a large share of the calls and a small share of the tokens — many short requests — or the reverse. Which one matters depends on whether your cost is driven by call volume or by prompt length. Measured on AIHubMix over the last 29 days, counting the 134 model IDs listed on this page; traffic routed through upstream-specific IDs that are not in the public catalog is not included.

All 134 OpenAI Models

Open in model list

OpenAI models on AIHubMix with input and output modalities, context length, maximum output, price per million tokens including cache read and cache write rates, and measured throughput and latency.
	Modalities
gpt-5.5-free	Takes text, vision, returns text.	1.05M	128K	FreeFree/M	Free/M	—	42 tok/s	3.58 s
gpt-5.6-luna	Takes text, vision, returns text.	1.05M	128K	$0.2$1.2/M	$0.02/M	$0.25/M	50 tok/s	4.75 s
gpt-5.6-terra	Takes text, vision, returns text.	1.05M	128K	$2$12/M	$0.2/M	$2.5/M	59 tok/s	3.61 s
gpt-5.4	Takes text, vision, returns text.	1.05M	128K	$2.5$15/M	$0.25/M	—	70 tok/s	1.69 s
gpt-5.4-high	Takes text, vision, returns text.	1.05M	128K	$2.5$15/M	$0.25/M	—	—	—
gpt-5.4-low	Takes text, vision, returns text.	1.05M	128K	$2.5$15/M	$0.25/M	—	—	—
gpt-5.6-sol	Takes text, vision, returns text.	1.05M	128K	$4$20/M	$0.4/M	$5/M	39 tok/s	6.31 s
gpt-5.6-sol-disc	Takes text, vision, returns text.	1.05M	128K	$4$20/M	$0.4/M	$5/M	37 tok/s	5.16 s
gpt-5.5	Takes text, vision, returns text.	1.05M	128K	$5$30/M	$0.5/M	—	64 tok/s	5.13 s
gpt-6-astra	Takes text, vision, returns text.	1.05M	128K	$10$50/M	$1/M	$12.5/M	21 tok/s	17.02 s
gpt-5.4-pro	Takes text, vision, returns text.	1.05M	128K	$30$180/M	—	—	42 tok/s	8.60 s
gpt-5.5-pro	Takes text, vision, returns text.	1.05M	128K	$30$180/M	—	—	14 tok/s	19.20 s
gpt-4.1-free	Takes text, vision, returns text.	1.05M	33K	FreeFree/M	Free/M	—	72 tok/s	0.53 s
gpt-4.1-mini-free	Takes text, vision, returns text.	1.05M	33K	FreeFree/M	Free/M	—	59 tok/s	0.36 s
gpt-4.1-nano-free	Takes text, vision, returns text.	1.05M	33K	FreeFree/M	Free/M	—	110 tok/s	0.33 s
gpt-4o-free	Takes text, vision, returns text.	1.05M	33K	FreeFree/M	Free/M	—	72 tok/s	0.53 s
gpt-4.1-nano	Takes text, vision, returns text.	1.05M	33K	$0.1$0.4/M	$0.025/M	—	73 tok/s	0.82 s
gpt-4.1-mini	Takes text, vision, returns text.	1.05M	33K	$0.4$1.6/M	$0.1/M	—	58 tok/s	0.74 s
gpt-4.1	Takes text, vision, returns text.	1.05M	33K	$2$8/M	$0.5/M	—	81 tok/s	1.13 s
auto	Takes text, vision, audio, video, returns text.	1M	—	FreeFree/M	—	—	—	—
gpt-5-nano	Takes text, vision, returns text.	400K	128K	$0.05$0.4/M	$0.005/M	—	87 tok/s	4.78 s
gpt-5.4-nano	Takes text, vision, returns text.	400K	128K	$0.2$1.25/M	$0.02/M	—	102 tok/s	0.65 s
gpt-5-mini	Takes text, vision, returns text.	400K	128K	$0.25$2/M	$0.025/M	—	94 tok/s	4.07 s
gpt-5.1-codex-mini	Takes text, vision, returns text.	400K	128K	$0.25$2/M	$0.025/M	—	168 tok/s	0.70 s
gpt-5.4-mini	Takes text, vision, returns text.	400K	128K	$0.75$4.5/M	$0.075/M	—	129 tok/s	1.07 s
gpt-5	Takes text, vision, returns text.	400K	128K	$1.25$10/M	$0.125/M	—	69 tok/s	7.76 s
gpt-5-chat-latest	Takes text, vision, returns text.	400K	128K	$1.25$10/M	$0.125/M	—	77 tok/s	0.85 s
gpt-5-codex	Takes text, vision, returns text.	400K	128K	$1.25$10/M	$0.125/M	—	64 tok/s	6.24 s
gpt-5.1	Takes text, vision, returns text.	400K	128K	$1.25$10/M	$0.125/M	—	71 tok/s	2.25 s
gpt-5.1-codex	Takes text, vision, returns text.	400K	128K	$1.25$10/M	$0.125/M	—	82 tok/s	0.62 s
gpt-5.1-codex-max	Takes text, vision, returns text.	400K	128K	$1.25$10/M	$0.125/M	—	—	—
gpt-5.2	Takes text, vision, returns text.	400K	128K	$1.75$14/M	$0.175/M	—	51 tok/s	1.36 s
gpt-5.2-codex	Takes text, vision, returns text.	400K	128K	$1.75$14/M	$0.175/M	—	82 tok/s	0.66 s
gpt-5.2-high	Takes text, vision, returns text.	400K	128K	$1.75$14/M	$0.175/M	—	—	—
gpt-5.2-low	Takes text, vision, returns text.	400K	128K	$1.75$14/M	$0.175/M	—	—	—
gpt-5.3-codex	Takes text, vision, returns text.	400K	128K	$1.75$14/M	$0.175/M	—	75 tok/s	11.73 s
gpt-chat-latest	Takes text, vision, returns text.	400K	128K	$5$30/M	$0.5/M	—	48 tok/s	1.93 s
gpt-5-pro	Takes text, vision, returns text.	400K	128K	$15$120/M	—	—	9 tok/s	313.10 s
gpt-5.2-pro	Takes text, vision, returns text.	400K	128K	$21$168/M	$2.1/M	—	9 tok/s	25.47 s
o3-mini	Takes text, vision, returns text.	200K	100K	$1.1$4.4/M	$0.55/M	—	474 tok/s	6.65 s
o4-mini	Takes text, vision, returns text.	200K	100K	$1.1$4.4/M	$0.275/M	—	86 tok/s	5.80 s
codex-mini-latest	Takes text, vision, returns text.	200K	—	$1.5$6/M	$0.375/M	—	—	—
o3	Takes text, vision, returns text.	200K	100K	$2$8/M	$0.5/M	—	40 tok/s	6.46 s
o3-pro	Takes text, vision, returns text.	200K	100K	$20$80/M	$20/M	—	13 tok/s	88.94 s
o1-pro	Takes text, returns text.	200K	—	$170$680/M	$170/M	—	19 tok/s	96.00 s
gpt-oss-20b-free	Takes text, returns text.	131K	—	FreeFree/M	—	—	—	—
gpt-oss-120b	Takes text, returns text.	131K	33K	$0.18$0.9/M	—	—	1101 tok/s	0.12 s
gpt-oss-20b	Takes text, returns text.	128K	—	$0.11$0.55/M	—	—	2619 tok/s	0.12 s
gpt-4o-mini	Takes text, vision, returns text.	128K	16K	$0.15$0.6/M	$0.075/M	—	49 tok/s	0.62 s
gpt-4o-mini-search-preview	Takes text, vision, returns text.	128K	16K	$0.15$0.6/M	$0.075/M	—	189 tok/s	1.57 s
gpt-5.1-chat-latest	Takes text, vision, returns text.	128K	16K	$1.25$10/M	$0.125/M	—	107 tok/s	0.94 s
gpt-5.2-chat-latest	Takes text, vision, returns text.	128K	16K	$1.75$14/M	$0.175/M	—	84 tok/s	0.77 s
gpt-5.3-chat-latest	Takes text, vision, returns text.	128K	16K	$1.75$14/M	$0.175/M	—	100 tok/s	0.82 s
gpt-4o	Takes text, vision, returns text.	128K	16K	$2.5$10/M	$1.25/M	—	52 tok/s	0.64 s
gpt-4o-2024-11-20	Takes text, vision, returns text.	128K	16K	$2.5$10/M	$1.25/M	—	61 tok/s	0.60 s
gpt-4o-audio-preview	Takes text, audio, returns text.	128K	16K	$2.5$10/M	—	—	10 tok/s	2.49 s
gpt-4o-search-preview	Takes text, vision, returns text.	128K	16K	$2.5$10/M	$1.25/M	—	121 tok/s	2.33 s
gpt-audio-1.5	Takes text, audio, returns text, audio.	128K	16K	$2.5$10/M	—	—	—	—
gpt-4o-2024-05-13		128K	4K	$5$15/M	$5/M	—	199 tok/s	0.45 s
gpt-4o-transcribe-diarize	Takes text, audio, returns text.	16K	2K	$2.5$10/M	—	—	—	—
o1	Takes text, returns text.	0K	—	$15$60/M	$7.5/M	—	39 tok/s	9.40 s
dall-e-2	Takes text, vision, returns vision.	—	—	FreeFree/M	—	—	—	—
dall-e-3	Takes text, vision, returns vision.	—	—	FreeFree/M	—	—	—	—
gpt-image-1	Takes text, vision, returns vision.	—	—	FreeFree/M	—	—	—	—
gpt-image-1-mini	Takes text, vision, returns vision.	—	—	FreeFree/M	—	—	—	—
gpt-image-1.5	Takes text, vision, returns vision.	—	—	FreeFree/M	—	—	—	—
gpt-image-2	Takes text, vision, returns vision.	—	—	FreeFree/M	Free/M	—	—	—
gpt-image-2-free	Takes text, vision, returns text, vision.	—	—	FreeFree/M	—	—	—	—
sora-2	Takes , returns video.	—	—	FreeFree/M	—	—	—	—
sora-2-pro	Takes , returns video.	—	—	Free$720/M	—	—	—	—
whisper-1	Takes audio, returns text.	—	—	FreeFree/M	—	—	—	—
whisper-large-v3	Takes audio, returns text.	—	—	FreeFree/M	—	—	—	—
whisper-large-v3-turbo	Takes audio, returns text.	—	—	FreeFree/M	—	—	—	—
omni-moderation-latest		—	—	$0.02$0.02/M	—	—	—	—
text-embedding-3-small	Takes text. Output modality not published.	—	—	$0.02$0.02/M	—	—	—	—
text-embedding-ada-002	Takes text. Output modality not published.	—	—	$0.1$0.1/M	—	—	—	—
text-embedding-v1	Takes text. Output modality not published.	—	—	$0.1$0.1/M	—	—	—	—
GPT-OSS-20B		—	—	$0.11$0.55/M	—	—	2619 tok/s	0.12 s
text-embedding-3-large	Takes text. Output modality not published.	—	—	$0.13$0.13/M	—	—	—	—
gpt-4o-mini-2024-07-18	Takes text, vision, returns text.	—	—	$0.15$0.6/M	$0.075/M	—	—	—
gpt-4o-mini-audio-preview	Takes text, audio, returns text.	—	—	$0.15$0.6/M	—	—	—	—
gpt-4o-mini-global		—	—	$0.15$0.6/M	$0.075/M	—	—	—
text-moderation-007		—	—	$0.2$0.2/M	—	—	—	—
text-moderation-latest		—	—	$0.2$0.2/M	—	—	—	—
text-moderation-stable		—	—	$0.2$0.2/M	—	—	—	—
aihubmix-router	Takes text, vision, returns text.	—	—	$0.4$1.6/M	$0.1/M	—	—	—
text-ada-001		—	—	$0.4$0.4/M	—	—	—	—
gpt-3.5-turbo		—	—	$0.5$1.5/M	—	—	—	—
text-babbage-001		—	—	$0.5$0.5/M	—	—	—	—
gpt-4o-mini-tts	Takes audio, returns audio.	—	—	$0.6$12/M	—	—	—	0.95 s
gpt-3.5-turbo-1106		—	—	$1$2/M	—	—	—	—
o3-mini-global		—	—	$1.1$4.4/M	$0.55/M	—	—	—
gpt-3.5-turbo-0301		—	—	$1.5$1.5/M	—	—	—	—
gpt-3.5-turbo-0613		—	—	$1.5$2/M	—	—	—	—
gpt-3.5-turbo-instruct		—	—	$1.5$2/M	—	—	—	—
davinci-002		—	—	$2$2/M	—	—	—	—
o3-global		—	—	$2$8/M	$0.5/M	—	—	—
text-curie-001		—	—	$2$2/M	—	—	—	—
gpt-4o-2024-08-06		—	—	$2.5$10/M	$1.25/M	—	—	—
gpt-4o-2024-08-06-global		—	—	$2.5$10/M	$1.25/M	—	—	—
gpt-4o-zh	Takes text, vision, returns text.	—	—	$2.5$10/M	—	—	—	—
computer-use-preview		—	—	$3$12/M	—	—	—	—
gpt-3.5-turbo-16k		—	—	$3$4/M	—	—	—	—
gpt-3.5-turbo-16k-0613		—	—	$3$4/M	—	—	—	—
o1-mini		—	—	$3$12/M	$1.5/M	—	—	—
o1-mini-2024-09-12		—	—	$3$12/M	$1.5/M	—	—	—
gpt-image-test		—	—	$5$40/M	—	—	—	—
distil-whisper-large-v3-en	Takes audio, returns text.	—	—	$5.556$5.556/M	—	—	—	—
gpt-4-0125-preview		—	—	$10$30/M	—	—	—	—
gpt-4-1106-preview		—	—	$10$30/M	—	—	—	—
gpt-4-turbo		—	—	$10$30/M	—	—	—	—
gpt-4-turbo-2024-04-09		—	—	$10$30/M	—	—	—	—
gpt-4-turbo-preview		—	—	$10$30/M	—	—	—	—
gpt-4-vision-preview		—	—	$10$30/M	—	—	—	—
o3-deep-research		—	—	$10$40/M	$2.5/M	—	—	—
o1-2024-12-17	Takes text, vision, returns text.	—	—	$15$60/M	$7.5/M	—	—	—
o1-preview	Takes text, vision, returns text.	—	—	$15$60/M	$7.5/M	—	—	—
o1-preview-2024-09-12		—	—	$15$60/M	$7.5/M	—	—	—
tts-1	Takes audio, returns audio.	—	—	$15$15/M	—	—	—	—
tts-1-1106	Takes audio, returns audio.	—	—	$15$15/M	—	—	—	—
davinci		—	—	$20$20/M	—	—	—	—
o3-pro-global		—	—	$20$80/M	—	—	—	—
text-davinci-002		—	—	$20$20/M	—	—	—	—
text-davinci-003		—	—	$20$20/M	—	—	—	—
text-davinci-edit-001		—	—	$20$20/M	—	—	—	—
text-search-ada-doc-001		—	—	$20$20/M	—	—	—	—
gpt-4		—	—	$30$60/M	—	—	—	—
gpt-4-0314		—	—	$30$60/M	—	—	—	—
gpt-4-0613		—	—	$30$60/M	—	—	—	—
tts-1-hd	Takes audio, returns audio.	—	—	$30$30/M	—	—	—	—
tts-1-hd-1106	Takes audio, returns audio.	—	—	$30$30/M	—	—	—	—
gpt-4-32k		—	—	$60$120/M	—	—	—	—
gpt-4-32k-0314		—	—	$60$120/M	—	—	—	—
gpt-4-32k-0613		—	—	$60$120/M	—	—	—	—

Prices are USD per million tokens; cache read and cache write are the rates for prompt-cache hits and for writing a prompt into the cache. Throughput and latency are measured on AIHubMix — the same figures the model detail page shows — not vendor claims. A dash means the catalog does not publish that field for that model, which is not the same as the model not supporting it.

Common Comparisons

aihubmix-router vs autoaihubmix-router vs codex-mini-latestaihubmix-router vs gpt-4.1aihubmix-router vs qwen3.8-max-previewaihubmix-router vs kimi-k3

OpenAI on AIHubMix

Which OpenAI model should I start with?

auto is free on input — the cheapest entry here that declares tool calling, and it carries a 1M context. Move up to o1-pro when answer quality matters more than cost, or to gpt-5.5-free for long-form reasoning.

Which of these models reason before answering?

43 of the 134 models here declare a reasoning phase — they work through the problem before producing an answer, which helps on multi-step problems at the cost of extra output tokens. Use the Reasoning filter above the table to see them. The catalog does not record anything further about how they differ, so this page does not sort them into families.

Why are there several entries for the same model?

Because each row is a route you can call, not a model release. Some IDs name an upstream (azure-, alicloud-, cc-), and some differ only in capitalisation, kept so older integrations keep working.

The catalog does not carry a field saying which of those a given row is, so this page does not sort them into buckets it would have to invent. Every row shows that route’s own price, context and speed — compare those directly, and open a model to see the upstreams that serve it.

How is cached input billed?

The Cache read column is the rate for input tokens served from the prompt cache — for example gpt-5 bills cache hits at 10% of the input rate and gpt-5-chat-latest bills cache hits at 10% of the input rate. Cache write is the surcharge for putting a prompt into the cache in the first place, and only a few upstreams bill it separately. A dash in either column means the catalog carries no cache rate for that model, so plan on paying the full input rate.

Do I need a separate OpenAI account?

No. One AIHubMix key covers every model on this page, and switching between them is a change to the model string — billing, rate limits, and logs stay in one place.