> Qwen3.8-2.4T-A95B is an open-weights AI model from Alibaba, released 12 Aug 2026. Estimated index score ~169 (GPT-4 = 100), not yet independently tested. API price $2.0 in / $6.0 out per million tokens, 1.05M context.

[Models](https://themodelindex.org/models/) /[Alibaba](https://themodelindex.org/labs/alibaba/)

# Qwen3.8-2.4T-A95B

Alibaba

Released 12 Aug 2026

Open

[Compare](https://themodelindex.org/compare/qwen3-8-2-4t-a95b-vs-gpt-6-astra/) [Run it yourself →](https://themodelindex.org/local/) [Announcement](https://huggingface.co/Qwen/Qwen3.8-2.4T-A95B)

~169158–181

Estimated index score

expected range · not ranked until independently tested

–

Core tests taken

no independent results yet

−17

Versus the frontier at release

best then: Claude Opus 5

$2.0/ $6.0

Price per million tokens

input / output, list price

1.05M

Context window

tokens it reads at once

## Where it sits · on the frontier

_Chart: Model Index score of every scored model since late 2022, with the best closed and best open model over time_

Dashed ring: an estimate until Epoch AI publishes results for this model.

## Test results · by area · tick = best by any model

Reasoning & knowledge

no evidence yet

GPQA Diamond

no result

Humanity's Last Exam

no result

Maths

no evidence yet

FrontierMath T1-3

no result

Coding

no evidence yet

SWE-bench Verified

no result

Terminal-Bench

no result

Agents

no evidence yet

APEX-Agents

no result

Novel problems

no evidence yet

ARC-AGI-2

no result

Long tasks

no evidence yet

METR time horizon

no result

Independent results from the Epoch AI benchmarking hub, best reasoning setting. The core tests are shown here; anchor tests also inform the score and can be shown above.

## What others say · rank on each leaderboard

Artificial Analysis Intelligence Index

overall capability across ten evaluations

#18

of 207 · 40

Further left is better. Leaderboards measure different things: LMArena is people voting blind between two answers, the others are test scores. Click any row for the source.

## Best at · top three among current models

Not in the top three among current models on any category we track.

Where this model places in the top three of today's models, on category leaderboards and on our core tests.

## Where to use it · prices checked 1 Oct 2026

#### From the lab

Qwen Chat

app · subscription

↗

Alibaba Cloud Model Studio

API

↗

#### API providers · 7 · per million tokens, in / out

N

Novita

$2.0 / $6.0

cheapest

A

Alibaba

$2.0 / $6.0

↗

S

SiliconFlow

$2.0 / $6.0 · fp8

↗

V

Venice

$2.0 / $6.0

↗

M

Modal

$2.0 / $6.0

↗

D

DeepInfra

$2.0 / $6.0 · fp4

↗

T

Together

$2.0 / $6.0

↗

#### Run it yourself

🤗

Official weights

Qwen/Qwen3.8-2.4T-A95B

↗

Provider prices come from OpenRouter's live listing; going direct to a provider can differ. Subscriptions are the lab's own apps.

## Facts and sources

Released

12 Aug 2026

Announcement

[huggingface.co ↗](https://huggingface.co/Qwen/Qwen3.8-2.4T-A95B)

Release entry

Qwen3.8-2.4T-A95B

Scored as

–

Artificial Analysis index

40

Output speed

40 tokens/s

Hugging Face

[Qwen/Qwen3.8-2.4T-A95B ↗](https://huggingface.co/Qwen/Qwen3.8-2.4T-A95B)

## Alibaba releases · around this one

Qwen3.7-Plus

Closed

1 JUN 2026

index 147

Qwen3.8-Max

Closed

3 AUG 2026

index 170

Qwen3.8-2.4T-A95B

Open

12 AUG 2026

~169 est.

Qwen3.8-27B

Open

14 AUG 2026

index 152

Qwen3.8-Flash-Next

Open

24 AUG 2026

~169 est.

---
Source: https://themodelindex.org/models/qwen3-8-2-4t-a95b/ · The Model Index · data as of 2026-10-01