Pass IndexThe State of AISign in

Qwen2.5 72B Instruct

Qwen2.5 is a model pretrained on a large-scale dataset of up to 18 trillion tokens, offering significant improvements in knowledge, coding, mathematics, and instruction following compared to its predecessor Qwen2.

text → text · made by Alibaba

$0.12$0.3per Mtok in / out
Hyperbolic · 12 sellers

Sold by 15 ways

SellerLaneRate
Hyperbolic aggregatorstandard$0.12per Mtok in$0.3per Mtok outraw.githubusercontent.com · read 2026-09-16
Requesty 2 ways$0.23per Mtok in$0.4per Mtok out$0.23per Mtok cached
Requesty aggregatordeepinfra$0.23per Mtok in$0.4per Mtok out$0.23per Mtok cachedrouter.requesty.ai · read 2026-08-28
Requesty aggregatornovita$0.38per Mtok in$0.4per Mtok out$0.38per Mtok cachedrouter.requesty.ai · read 2026-08-28
DeepInfra aggregatorstandard$0.36per Mtok in$0.4per Mtok outapi.deepinfra.com · read 2026-08-25
Hugging Face Inference Providers 2 ways$0.36per Mtok in$0.4per Mtok out
Hugging Face Inference Providers aggregatordeepinfra$0.36per Mtok in$0.4per Mtok outrouter.huggingface.co · read 2026-08-25
Hugging Face Inference Providers aggregatornovita$0.38per Mtok in$0.4per Mtok outrouter.huggingface.co · read 2026-08-25
Nous Research aggregatorstandard$0.36per Mtok in$0.4per Mtok outinference-api.nousresearch.com · read 2026-09-12
OpenRouter 2 ways$0.36per Mtok in$0.4per Mtok out
OpenRouter aggregatorstandard$0.36per Mtok in$0.4per Mtok outopenrouter.ai · read 2026-08-24
OpenRouter aggregatornovita/bf16$0.38per Mtok in$0.4per Mtok outopenrouter.ai · read 2026-08-25
Novita AI aggregatorstandard$0.38per Mtok in$0.4per Mtok outapi.novita.ai · read 2026-08-25
SiliconFlow aggregatorstandard$0.59per Mtok in$0.59per Mtok outmodels.dev · read 2026-09-16
Fireworks AI aggregatorstandard$0.9per Mtok in$0.9per Mtok outraw.githubusercontent.com · read 2026-09-16
ElectronHub aggregatorstandard$1per Mtok in$1per Mtok outapi.electronhub.ai · read 2026-09-16
Together AI aggregatorstandard$1.2per Mtok in$1.2per Mtok outraw.githubusercontent.com · read 2026-09-16
PPIO aggregatorstandard$4per Mtok in$12per Mtok outapi.ppinfra.com · read 2026-09-16

Measured 14 standings

PlaceBoardMetricScore
10thof 217Mmlu — Epoch AIEM0.853
12thof 16The agent company — Epoch AI% Score0.118
18thof 25AgentBench Leaderboard (new)Success Rate (pass@1) AVG40.8
28thof 36Balrog — Epoch AIAverage progress0.162
40thof 50Metr time horizons — Epoch AIaverage_score0.358
52ndof 108Math level 5 — Epoch AImean_score63.17
63rdof 81Forecastbench — Epoch AIOverall score57.5
103rdof 161Dtbench — Epoch AIAccuracy62.93
105thof 123Lmca — Epoch AIScore13.42
159thof 170Weirdml — Epoch AIAccuracy15.97
231stof 313GPQA Diamondmean_score49.148
242ndof 290OTIS Mock AIME 2024-2025mean_score8.056
272ndof 402LMArena · TextScore (Elo)1303
530thof 656Epoch capabilities index — Epoch AIECI Score129

About

Qwen2.5 72B Instruct — a text model from Alibaba, sold by 12 companies from $0.12 in and $0.3 out per million tokens, placed 10th of 217 on Mmlu — Epoch AI.

It takes text and returns text, with a context window of 32,768 tokens. It was published in September 2024, trained on material up to April 2024. Its sellers say it can call a tool. The catalogue files it under reasoning and agents. Twelve companies sell it. The cheapest is $0.12 in and $0.3 out per million tokens at Hyperbolic. Beside the standard rate there are separately routed and regional lanes. It has been measured on 14 boards, and stands best at 10th of 217 on Mmlu — Epoch AI.

Every current figure

Maker
Alibaba
Register
model
Takes
text
Returns
text
Context
32,768 tokens
Longest answer
16,384 tokens
Published
September 2024
Knowledge to
April 2024
Parameters
73 billion
Licence
not read
Sellers
12
Maker's own price
not read
Price
$0.12 in and $0.3 out per million tokens — Hyperbolic
Boards
14
Best place
10th of 217 — Mmlu — Epoch AI

Known as 5 names

Qwen2.5-72B-InstructQwen/Qwen2.5-72B-Instructqwen/qwen-2.5-72b-instructdeepinfra/Qwen/Qwen2.5-72B-Instructnovita/qwen/qwen-2.5-72b-instruct