Pass IndexThe State of AISign in

Qwen3.5-122B-A10B

Qwen3.5-122B-A10B is a large Mixture-of-Experts model from Alibaba's Qwen3.5 series with 122B total parameters and 10B activated per token.

text + image + video → text · made by Alibaba

$0.22$2.2per Mtok in / out
Databricks · 8 sellers

Sold by 15 ways

SellerLaneRate
Databricks cloudstandard$0.22per Mtok in$2.2per Mtok out$0.22per Mtok cached$0.22per Mtok cache writeraw.githubusercontent.com · read 2026-09-16
Nous Research aggregatorstandard$0.26per Mtok in$2.08per Mtok outinference-api.nousresearch.com · read 2026-09-12
OpenRouter 6 ways$0.26per Mtok in$2.08per Mtok out
OpenRouter aggregatorstandard$0.26per Mtok in$2.08per Mtok outopenrouter.ai · read 2026-09-10
OpenRouter aggregatoralibaba$0.26per Mtok in$2.08per Mtok outopenrouter.ai · read 2026-08-29
OpenRouter aggregatorsiliconflow/fp8$0.26per Mtok in$2.08per Mtok outopenrouter.ai · read 2026-08-30
OpenRouter aggregatordeepinfra/fp4$0.29per Mtok in$2.4per Mtok outopenrouter.ai · read 2026-08-25
OpenRouter aggregatoratlas-cloud/fp8$0.3per Mtok in$2.4per Mtok out$0.3per Mtok cachedopenrouter.ai · read 2026-08-25
OpenRouter aggregatornovita/bf16$0.4per Mtok in$3.2per Mtok outopenrouter.ai · read 2026-08-25
SiliconFlow aggregatorstandard$0.26per Mtok in$2.08per Mtok outmodels.dev · read 2026-09-16
DeepInfra aggregatorstandard$0.29per Mtok in$2.4per Mtok outapi.deepinfra.com · read 2026-08-25
Hugging Face Inference Providers 3 ways$0.4per Mtok in$3.2per Mtok out
Hugging Face Inference Providers aggregatordeepinfra$0.29per Mtok in$2.4per Mtok outrouter.huggingface.co · read 2026-08-25
Hugging Face Inference Providers aggregatorstandard$0.4per Mtok in$3.2per Mtok outmodels.dev · read 2026-09-16
Hugging Face Inference Providers aggregatornovita$0.4per Mtok in$3.2per Mtok outrouter.huggingface.co · read 2026-08-25
Novita AI aggregatorstandard$0.4per Mtok in$3.2per Mtok outapi.novita.ai · read 2026-08-25
PPIO aggregatorstandard$2per Mtok in$16per Mtok outapi.ppinfra.com · read 2026-09-16

Measured 11 standings

PlaceBoardMetricScore
21stof 262MMLU-ProOverall (accuracy)0.867
48thof 161Dtbench — Epoch AIAccuracy (none)84.27
60thof 123Lmca — Epoch AIScore (none)32.23
66thof 127Mystery game puzzles — Epoch AImean_score (none)17
71stof 105Vectara Hallucination LeaderboardHallucination Rate11.2
95thof 143Artificial Analysis · Agentic IndexIndex9.6
101stof 184Artificial Analysis · Coding IndexIndex45.7
118thof 177Critpt — Epoch AIAccuracy (none)0.857
130thof 302Artificial Analysis · Intelligence IndexArtificial Analysis Intelligence Index18
139thof 168Scicode — Epoch AIScore (none)35.648
292ndof 656Epoch capabilities index — Epoch AIECI Score (none)146.26

About

Qwen3.5-122B-A10B — a text model from Alibaba, sold by 8 companies from $0.22 in and $2.2 out per million tokens, placed 21st of 262 on MMLU-Pro.

It takes text, images and video and returns text, with a context window of 262,144 tokens. It was published in February 2026, trained on material up to April 2025. Its sellers say it can reason step by step and call a tool. The catalogue files it under extract and reasoning. Eight companies sell it. The cheapest is $0.22 in and $2.2 out per million tokens at Databricks. Beside the standard rate there are separately routed and regional lanes. It has been measured on 11 boards, and stands best at 21st of 262 on MMLU-Pro.

Every current figure

Maker
Alibaba
Register
model
Takes
text + image + video
Returns
text
Context
262,144 tokens
Published
February 2026
Knowledge to
April 2025
Parameters
125 billion
Licence
not read
Sellers
8
Maker's own price
not read
Price
$0.22 in and $2.2 out per million tokens — Databricks
Boards
11
Best place
21st of 262 — MMLU-Pro

Known as 1 name

Qwen/Qwen3.5-122B-A10B