AIHubMixAIHubMix

Microsoft Models

16 modelsGeneral models free to startUp to 256K context

Usage

Last 28 days · 2026-08-10 to 2026-09-07

Tokens

20.9K

Requests

594

Models in use

8 of 16

Tokens per day, stacked by model

06K11.9K08-1008-1808-2509-012026-08-10 — 0 tokens2026-08-11 — 0 tokens2026-08-12 — 0 tokens2026-08-14 — 0 tokens2026-08-15 — 1,425 tokens mai-thinking-1: 1,4252026-08-16 — 0 tokens2026-08-17 — 0 tokens2026-08-18 — 0 tokens2026-08-19 — 1,640 tokens mai-thinking-1: 1,6402026-08-20 — 0 tokens2026-08-21 — 0 tokens2026-08-22 — 1,460 tokens mai-thinking-1: 1,4602026-08-23 — 0 tokens2026-08-24 — 0 tokens2026-08-25 — 0 tokens2026-08-26 — 0 tokens2026-08-27 — 0 tokens2026-08-28 — 725 tokens mai-thinking-1: 7252026-08-29 — 0 tokens2026-08-30 — 0 tokens2026-08-31 — 1,590 tokens mai-thinking-1: 1,5902026-09-01 — 955 tokens mai-thinking-1: 9552026-09-02 — 335 tokens mai-thinking-1: 3352026-09-03 — 0 tokens2026-09-04 — 0 tokens2026-09-05 — 0 tokens2026-09-06 — 905 tokens mai-thinking-1: 9052026-09-07 — 11,910 tokens mai-thinking-1: 11,910

Which models that traffic went to

  1. Mai Thinking 1100.0%20.9K

Share of 20.9K tokens. 7 models with traffic report no token counts and cannot be ranked here, including mai-image-2.5-pro and mai-image-2.5-flash — 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 28 days, counting the 16 model IDs listed on this page; traffic routed through upstream-specific IDs that are not in the public catalog is not included.

All 16 Microsoft Models

Open in model list
Microsoft 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
mai-thinking-1Takes text, returns text.256K—$2$8/M42 tok/s18.97 s
aihub-Phi-4-mini-instructTakes text, returns text.128K4K$0.12$0.48/M——
aihub-Phi-4-multimodal-instructTakes text, vision, audio, returns text.128K4K$0.12$0.48/M——
AiHubmix-Phi-4-mini-reasoningTakes text, returns text.128K4K$0.12$0.12/M——
mai-image-2.6Takes text, vision, returns text, vision.32K—FreeFree/M——
mai-image-2.6-flashTakes text, vision, returns text, vision.32K—FreeFree/M——
aihub-Phi-4Takes text, returns text.16K—$0.12$0.48/M46 tok/s0.41 s
mai-image-2.5Takes text, vision, returns text, vision.——FreeFree/M——
mai-image-2.5-flashTakes text, vision, returns text, vision.——FreeFree/M——
mai-image-2.5-proTakes text, vision, returns text, vision.——FreeFree/M——
ahm-Phi-3-5-MoE-instruct——$0.4$1.6/M——
ahm-Phi-3-5-vision-instructTakes text, vision, returns text.——$0.4$1.6/M——
ahm-Phi-3-5-mini-instruct——$1$3/M——
ahm-Phi-3-medium-4k——$1$3/M——
ahm-Phi-3-small-128k——$1$3/M——
ahm-Phi-3-medium-128k——$6$18/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-Phi-4-mini-reasoning vs ahm-Phi-3-5-vision-instructAiHubmix-Phi-4-mini-reasoning vs aihub-Phi-4AiHubmix-Phi-4-mini-reasoning vs aihub-Phi-4-mini-instructAiHubmix-Phi-4-mini-reasoning vs qwen3.8-max-previewAiHubmix-Phi-4-mini-reasoning vs kimi-k3

Microsoft on AIHubMix

Which Microsoft model should I start with?

mai-image-2.5 is free on input — the cheapest entry here that declares a token price. Move up to ahm-Phi-3-medium-128k when answer quality matters more than cost, or to mai-thinking-1 for long-form reasoning.

Which of these models reason before answering?

1 of the 16 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.

Do I need a separate Microsoft 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.

Other Model Authors

See all model authors →
Qwen
144 modelsOpenAI
134 modelsGoogle
85 modelsZ.AI
71 modelsLlama
53 models

Start calling Microsoft in one line

One key, one endpoint, 880 models across 38 model authors.

Get API keyRead the docs
AIHubMix

© 2023 - 2026 AIHubMix, LLC

Product

Company

Developer

Connect