AIHubMixAIHubMix

Grok Models

25 modelsGeneral models from $0.20/M inputUp to 2M context

Usage

Last 30 days · 2026-07-13 to 2026-08-11

Tokens

36.8B

Requests

359K

Models in use

19 of 25

Tokens per day

0945M1.9B07-1307-2007-2708-0308-102026-07-13 — 1,601,219,625 tokens2026-07-14 — 1,065,560,835 tokens2026-07-15 — 1,815,782,590 tokens2026-07-16 — 1,772,491,450 tokens2026-07-17 — 1,026,366,870 tokens2026-07-18 — 588,775,920 tokens2026-07-19 — 629,320,935 tokens2026-07-20 — 1,415,549,770 tokens2026-07-21 — 1,182,214,775 tokens2026-07-22 — 1,282,063,440 tokens2026-07-23 — 1,601,836,420 tokens2026-07-24 — 1,602,372,145 tokens2026-07-25 — 553,992,020 tokens2026-07-26 — 1,045,837,685 tokens2026-07-27 — 1,345,329,550 tokens2026-07-28 — 1,693,856,335 tokens2026-07-29 — 1,433,916,755 tokens2026-07-30 — 1,032,333,625 tokens2026-07-31 — 754,770,980 tokens2026-08-01 — 781,478,030 tokens2026-08-02 — 1,078,064,620 tokens2026-08-03 — 1,119,451,225 tokens2026-08-04 — 1,596,586,665 tokens2026-08-05 — 1,064,676,540 tokens2026-08-06 — 1,249,699,005 tokens2026-08-07 — 1,237,580,490 tokens2026-08-08 — 637,345,900 tokens2026-08-09 — 895,444,215 tokens2026-08-10 — 1,889,068,535 tokens2026-08-11 — 1,770,328,060 tokens

Which models that traffic went to

  1. grok-4.570.9%26.1B
  2. grok-4-1-fast-non-reasoning18.4%6.8B
  3. grok-4-20-non-reasoning3.7%1.4B
  4. grok-4-1-fast-reasoning3.6%1.3B
  5. grok-4.33.0%1.1B
  6. grok-4-20-reasoning0.2%66.1M
  7. grok-40.1%26.9M
  8. grok-4-fast-non-reasoning0.1%19.3M
  9. 5 more models0.1%29M

Share of 36.8B tokens. 6 models with traffic report no token counts and cannot be ranked here, including grok-3-mini and grok-4.20-multi-agent-beta-0309 — 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 30 days, counting the 25 model IDs listed on this page; traffic routed through upstream-specific IDs that are not in the public catalog is not included.

All 25 Grok Models

Open in model list
Grok 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
grok-4-1-fast-non-reasoningTakes text, vision, returns text.2M—$0.20$0.50/M$0.05/M53 tok/s1.30 s
grok-4-1-fast-reasoningTakes text, vision, returns text.2M—$0.20$0.50/M$0.05/M68 tok/s3.53 s
grok-4-fast-non-reasoningTakes text, vision, returns text.2M30K$0.20$0.50/M$0.05/M137 tok/s2.70 s
grok-4-fast-reasoningTakes text, vision, returns text.2M30K$0.20$0.50/M$0.05/M137 tok/s2.70 s
grok-4-20-non-reasoningTakes text, vision, returns text.2M—$2.00$6.00/M$0.20/M112 tok/s1.51 s
grok-4-20-reasoningTakes text, vision, returns text.2M—$2.00$6.00/M$0.20/M11 tok/s1.88 s
grok-4.20-beta-0309-non-reasoningTakes text, vision, returns text.2M—$2.00$6.00/M$0.20/M89 tok/s0.91 s
grok-4.20-beta-0309-reasoningTakes text, vision, returns text.2M—$2.00$6.00/M$0.20/M74 tok/s1.69 s
grok-4.20-multi-agent-0309Takes text, vision, returns text.2M—$2.00$6.00/M$0.20/M43 tok/s4.26 s
grok-4.20-multi-agent-beta-0309Takes text, vision, returns text.2M—$2.00$6.00/M$0.20/M43 tok/s4.26 s
grok-4.3Takes text, vision, returns text.1M—$1.25$2.50/M$0.20/M17 tok/s0.85 s
grok-4.5Takes text, vision, returns text.500K—$2.00$6.00/M$0.50/M42 tok/s2.00 s
grok-code-fast-1Takes text, vision, returns text.256K10K$0.20$0.50/M$0.05/M87 tok/s1.26 s
grok-build-0.1Takes text, vision, returns text.256K—$1.00$2.00/M$0.20/M7 tok/s0.38 s
grok-4Takes text, vision, returns text.256K64K$3.30$16.50/M$0.82/M54 tok/s2.70 s
grok-3-mini——$0.30$0.50/M———
grok-3-mini-beta——$0.33$0.55/M———
grok-3-mini-fast-beta——$0.33$2.20/M———
grok-2-1212——$1.80$9.00/M———
grok-2-vision-1212Takes text, vision, returns text.——$1.80$9.00/M———
grok-3——$3.00$15.00/M———
grok-3-beta——$3.00$15.00/M———
grok-3-fast——$5.50$27.50/M———
grok-3-fast-beta——$5.50$27.50/M———
grok-vision-betaTakes text, vision, returns text.——$5.60$16.80/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

grok-2-vision-1212 vs grok-4grok-2-vision-1212 vs grok-4-1-fast-non-reasoninggrok-2-vision-1212 vs grok-4-1-fast-reasoninggrok-2-vision-1212 vs qwen3.8-max-previewgrok-2-vision-1212 vs kimi-k3

Grok on AIHubMix

Which Grok model should I start with?

grok-4-1-fast-non-reasoning at $0.20/M input — the cheapest entry here that declares tool calling, and it carries a 2M context. Move up to grok-vision-beta when answer quality matters more than cost, or to grok-4-1-fast-reasoning for long-form reasoning.

Which of these models reason before answering?

13 of the 25 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 grok-4-20-non-reasoning bills cache hits at 10% of the input rate and grok-4-20-reasoning 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 Grok 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 Providers

Qwen
140 modelsOpenAI
132 modelsGoogle
81 modelsZ.AI
64 modelsLlama
52 models

Start calling Grok in one line

One key, one endpoint, 844 models across 35 providers.

Get API keyRead the docs
AIHubMix

© 2023 - 2026 AIHubMix, LLC

Product

Company

Developer

Connect