Meta Models

4 modelsGeneral models from $1.375/M inputUp to 1.05M context

Usage

Last 30 days · 2026-10-23 to 2026-11-21

Tokens

2.3B

Requests

30.9K

Models in use

4 of 4

Tokens per day, stacked by model

088.6M177M10-2310-3011-0611-1311-202026-10-23 — 7,971,980 tokens muse-spark-1.3: 5,450,965 muse-wonder: 2,494,160 muse-spark-1.1: 13,520 muse-spark-1.2: 13,3352026-10-24 — 45,052,065 tokens muse-spark-1.3: 30,017,630 muse-wonder: 15,019,210 muse-spark-1.1: 7,620 muse-spark-1.2: 7,6052026-10-25 — 97,764,835 tokens muse-spark-1.3: 95,119,865 muse-wonder: 2,631,420 muse-spark-1.1: 7,310 muse-spark-1.2: 6,2402026-10-26 — 84,318,230 tokens muse-spark-1.3: 74,636,875 muse-wonder: 9,656,040 muse-spark-1.1: 13,220 muse-spark-1.2: 12,0952026-10-27 — 74,325,160 tokens muse-spark-1.3: 68,410,250 muse-wonder: 5,898,305 muse-spark-1.2: 8,360 muse-spark-1.1: 8,2452026-10-28 — 84,632,215 tokens muse-spark-1.3: 80,159,585 muse-wonder: 4,457,225 muse-spark-1.2: 7,920 muse-spark-1.1: 7,4852026-10-29 — 61,303,345 tokens muse-spark-1.3: 39,087,190 muse-wonder: 22,200,260 muse-spark-1.2: 8,420 muse-spark-1.1: 7,4752026-10-30 — 81,574,970 tokens muse-spark-1.3: 47,164,870 muse-wonder: 34,387,060 muse-spark-1.1: 11,565 muse-spark-1.2: 11,4752026-10-31 — 108,868,245 tokens muse-spark-1.3: 106,912,820 muse-wonder: 1,928,040 muse-spark-1.1: 13,715 muse-spark-1.2: 13,6702026-11-01 — 103,359,230 tokens muse-spark-1.3: 94,425,995 muse-wonder: 8,917,770 muse-spark-1.1: 7,945 muse-spark-1.2: 7,5202026-11-02 — 63,347,985 tokens muse-spark-1.3: 42,435,155 muse-wonder: 20,893,150 muse-spark-1.2: 11,000 muse-spark-1.1: 8,6802026-11-03 — 78,571,030 tokens muse-spark-1.3: 56,816,050 muse-wonder: 21,726,270 muse-spark-1.2: 20,735 muse-spark-1.1: 7,9752026-11-04 — 103,884,345 tokens muse-spark-1.3: 80,301,325 muse-wonder: 23,572,975 muse-spark-1.2: 5,225 muse-spark-1.1: 4,8202026-11-05 — 116,637,190 tokens muse-spark-1.3: 111,062,255 muse-wonder: 5,558,535 muse-spark-1.1: 8,710 muse-spark-1.2: 7,6902026-11-06 — 138,455,320 tokens muse-spark-1.3: 130,780,145 muse-wonder: 7,656,535 muse-spark-1.1: 9,735 muse-spark-1.2: 8,9052026-11-07 — 27,146,505 tokens muse-spark-1.3: 27,123,770 muse-spark-1.1: 8,295 muse-spark-1.2: 7,345 muse-wonder: 7,0952026-11-08 — 66,656,925 tokens muse-spark-1.3: 66,574,470 muse-wonder: 64,950 muse-spark-1.2: 9,365 muse-spark-1.1: 8,1402026-11-09 — 57,744,335 tokens muse-spark-1.3: 56,529,090 muse-wonder: 1,183,700 muse-spark-1.1: 16,605 muse-spark-1.2: 14,9402026-11-10 — 177,208,095 tokens muse-spark-1.3: 176,971,465 muse-wonder: 221,615 muse-spark-1.1: 9,105 muse-spark-1.2: 5,9102026-11-11 — 72,053,135 tokens muse-spark-1.3: 54,428,590 muse-wonder: 17,600,290 muse-spark-1.2: 12,480 muse-spark-1.1: 11,7752026-11-12 — 126,534,355 tokens muse-spark-1.3: 73,928,905 muse-wonder: 52,591,490 muse-spark-1.1: 7,375 muse-spark-1.2: 6,5852026-11-13 — 36,541,230 tokens muse-spark-1.3: 36,182,925 muse-wonder: 342,350 muse-spark-1.1: 8,370 muse-spark-1.2: 7,5852026-11-14 — 57,745,360 tokens muse-spark-1.3: 36,459,305 muse-wonder: 21,270,760 muse-spark-1.1: 7,820 muse-spark-1.2: 7,4752026-11-15 — 120,838,500 tokens muse-spark-1.3: 86,151,395 muse-wonder: 34,666,700 muse-spark-1.2: 11,065 muse-spark-1.1: 9,3402026-11-16 — 68,748,280 tokens muse-spark-1.3: 61,715,255 muse-wonder: 7,019,005 muse-spark-1.1: 7,885 muse-spark-1.2: 6,1352026-11-17 — 82,457,295 tokens muse-spark-1.3: 63,243,980 muse-wonder: 19,193,370 muse-spark-1.1: 11,035 muse-spark-1.2: 8,9102026-11-18 — 34,472,645 tokens muse-spark-1.3: 23,140,335 muse-wonder: 11,313,885 muse-spark-1.2: 10,215 muse-spark-1.1: 8,2102026-11-19 — 51,135,695 tokens muse-spark-1.3: 33,972,840 muse-wonder: 17,146,775 muse-spark-1.1: 8,080 muse-spark-1.2: 8,0002026-11-20 — 60,377,585 tokens muse-spark-1.3: 60,312,815 muse-wonder: 48,760 muse-spark-1.1: 8,305 muse-spark-1.2: 7,7052026-11-21 — 50,678,135 tokens muse-spark-1.3: 31,060,930 muse-wonder: 19,600,960 muse-spark-1.1: 8,570 muse-spark-1.2: 7,675
  • muse-spark-1.3
  • muse-wonder
  • muse-spark-1.2
  • muse-spark-1.1

Which models that traffic went to

  1. Muse Spark 1.383.3%2B
  2. Muse Wonder16.6%389M
  3. Muse Spark 1.2<0.1%282K
  4. Muse Spark 1.1<0.1%277K

Share of 2.3B tokens.

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 4 model IDs listed on this page; traffic routed through upstream-specific IDs that are not in the public catalog is not included.

All 4 Meta Models

Open in model list
Meta 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
muse-spark-1.1Takes text, vision, audio, video, PDF, returns text.1.05M—$1.375$4.675/M———
muse-spark-1.2Takes text, vision, audio, video, PDF, returns text.1.05M—$1.375$4.675/M—108 tok/s14.25 s
muse-spark-1.3Takes text, vision, audio, video, PDF, returns text.1.05M1M$1.375$4.675/M$0.165/M143 tok/s2.96 s
muse-wonderTakes text, vision, audio, video, PDF, returns text.1.05M—$4.4$22/M$0.22/M75 tok/s12.88 s

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.

Meta on AIHubMix

Which Meta model should I start with?

muse-spark-1.1 at $1.375/M input — the cheapest entry here that declares tool calling, and it carries a 1.05M context. Move up to muse-wonder when answer quality matters more than cost, or to muse-spark-1.2 for long-form reasoning.

Which of these models reason before answering?

3 of the 4 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 muse-wonder bills cache hits at 5% of the input rate and muse-spark-1.3 bills cache hits at 12% 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 Meta 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.

Start calling Meta in one line

One key, one endpoint, 984 models across 49 model authors.