THE MODEL, IN DETAIL / alibaba

qwen3-235b-a22b-thinking-2507

qwen/qwen3-235b-a22b-thinking-2507Open weightsReasoning

Qwen3-235B-A22B-Thinking-2507 is a high-performance, open-weight Mixture-of-Experts (MoE) language model optimized for complex reasoning tasks. It activates 22B of its 235B parameters per forward pass and natively supports up to 262,144...

OpenRouter

Source-linked model intelligenceCatalog added 25 Jul 2025Specifications observed 17 Oct 2026

Text models

#131

Arena rating

Context window

128K

16.4K max output tokens

Input price

$0.45

USD / million input tokens

Output price

$3.5

USD / million output tokens

Inference providers

1 provider routes observed

Measured output speed

—

No model-level measurement

01 / PERFORMANCE

A clearer view of capability.

Read the score, understand the context, and explore where this model sits among its peers.

Performance fingerprint

Individual published accuracy scores · not a composite index

At least three published accuracy scores are needed for this view.

Performance meets price

Text models · same published source and version

GoogleAnthropicMetaxiaomi

In good company

Text models · Arena · official dataset

1,532.3

Google

1,515.8

Anthropic

1,511.6

Anthropic

1,507.6

Anthropic

1,502.5

Anthropic

1,501.7

Anthropic

1,501

Evidence, by task.

Coverage is not a capability score.

0published results
0 evaluation sources

ReasoningAwaiting evidence

CodingAwaiting evidence

MathematicsAwaiting evidence

VisionAwaiting evidence

AgenticAwaiting evidence

Long contextAwaiting evidence

Documented features, task evaluations, and preference rankings describe different things. Missing coverage does not imply a model cannot perform a task.

Explore our methodology

Where people rank it

Published rankings for this exact model or configuration. Each source keeps its original metric.

Text modelsArena rating · 15 Oct 2026#1311,415Confidence interval: 1,408.5–1,421.6

02 / BENCHMARK EXPLORER

The numbers behind the model.

Original results, exact versions, and a direct path to the evidence behind every score.

0 published results

Percentage-based evaluations will appear here when published.

0 percentage-based evaluations. Each benchmark measures a different task; compare results within the same evaluation and protocol.

Artificial Analysis indices

Published in the OpenRouter catalog · source-defined index values

Coding Index22.1

Agentic Index1.3

Intelligence Index12.7

Catalog observation: 17 Oct 2026. Index versions and evaluation settings are not supplied in this feed. These are not Singularity composite scores.

Artificial Analysis

Design Arena

Source-reported Elo across individual design categories

3d998

models arena · Rank #109 · 40.4% win rate

Codecategories1,034

models arena · Rank #116 · 40.8% win rate

Dataviz950

models arena · Rank #119 · 32.6% win rate

Gamedev960

models arena · Rank #124 · 34.2% win rate

Uicomponent944

models arena · Rank #119 · 34% win rate

Website1,054

models arena · Rank #116 · 42% win rate

Catalog observation: 17 Oct 2026. Category ratings have separate comparison pools.

03 / COST & VALUE

Know what your workload costs.

Explore token prices, see where the spend goes, and model a workload of your own.

Input$0.45per million tokens

Output$3.5per million tokens

Cache read—per million tokens

Cache write—per million tokens

Workload cost calculator

Change the numbers. See the trade-offs.

Pricing source

Input tokens / requestOutput tokens / requestRequests

Cached share of input0%

Cache-read pricing is not documented for this selection.

Token estimate only. Excludes cache writes, tools, and other provider fees. Actual token usage and provider limits vary.

Your estimated spend

1,000 requests · 11M tokens

Uncached input$4.5

Cached input$0

Output$3.5

Per request$0.008

Without cache discount$8

Estimated cache savings$0

04 / INFERENCE NETWORK

One model. Different ways to run it.

Compare inference routes by price, context, quantization, and reported availability.

1 providers · 1 routes

Standard service tier only

Provider routes for qwen3-235b-a22b-thinking-2507, observed 17 Oct 2026
Provider / route	Input / M	Output / M	Cache / M	Context	Output limit	Speed	Latency	Availability · 1d	Rates
VVeniceFP8	$0.45	$3.5	—	128K	16.4K	—	—	100%

Snapshot observed 17 Oct 2026. Speed and latency are the reported 30-minute median when available. This is a dated snapshot, not live monitoring.

Provider price comparison

USD per million tokens · input and output rates

InputOutput

Venicefp8

$0.45

$3.5

Reported availability

Provider availability over the selected observation window

Venice · FP8100%

Output speed

Reported 30-minute p50 · tokens per second

No provider throughput measurements were returned in this snapshot.

Response latency

Reported 30-minute p50 · seconds

No provider latency measurements were returned in this snapshot.

Context by provider

Actual route limits can differ from the catalog maximum

Venice · FP8128K

05 / CAPABILITIES & MODALITIES

What goes in. What comes out.

A map of documented interfaces and API features, separate from benchmark performance.

INPUT MODALITIES

Text

qwen3-235b-a22b-thinking-2507Reasoning supported

OUTPUT MODALITIES

Text

Feature support

Listed catalog API parameters and documented pricing features

ReasoningSupported

Tool callingSupported

Tool choiceSupported

Structured outputsNot listed

Response formatNot listed

Reasoning detailsSupported

Seed controlNot listed

Stop sequencesSupported

Prompt cachingNot documented

TemperatureSupported

Room to work

Context capacity and completion limits

128Ktotal context tokens

Context window

16.4Kmax output

Output tokens share the total context budget. The maximum completion length is a separate limit, and individual provider routes may have smaller limits.

Catalog context: 128,000 tokens
Maximum output: 16,384 tokens
Top-provider context: 128,000 tokens

Features vary by route

Check the provider before depending on a specific API feature.

API feature support by provider
Provider route	Reasoning	Tools	Structured output	Response format	Implicit caching
Venice · FP8	Supported	Supported	Not listed	Not listed	Not listed

06 / UNDER THE HOOD

The complete specification sheet.

Exact identities, model architecture, limits, and the defaults that shape a request.

Identity & architecture

Developer: alibaba
Exact identity: qwen3-235b-a22b-thinking-2507
Subject: Model Configuration
Access: Open weights
License: Apache 2.0
Parameters: Not disclosed
Tokenizer: Qwen3
Instruction format: qwen3

Reasoning & controls

Reasoning: Supported
Mandatory reasoning: Yes
Reasoning by default: Not documented
Published configuration: Not disclosed
Supported reasoning efforts: Not documented
Default reasoning effort: Not disclosed
API parameters: 11
Provider moderation: No
Quantization variants: fp8

Dates & limits

Catalog added: 25 Jul 2025
Knowledge cutoff: 2025-06-30
Catalog expiration: Not documented
Specifications observed: 17 Oct 2026
Providers observed: 17 Oct 2026
Request limits: Provider-specific / not supplied
Canonical catalog ID: qwen/qwen3-235b-a22b-thinking-2507

API parameters & defaults

Availability is catalog-wide. Defaults can vary across providers.

frequency_penaltySupported · default not supplied

include_reasoningSupported · default not supplied

max_tokensSupported · default not supplied

presence_penaltySupported · default not supplied

reasoningSupported · default not supplied

stopSupported · default not supplied

temperatureSupported · default not supplied

tool_choiceSupported · default not supplied

toolsSupported · default not supplied

top_kSupported · default not supplied

top_pSupported · default not supplied

07 / BUILD WITH THIS MODEL

From exploration to your first request.

Use the documented catalog identifier and keep your API key in an environment variable.

Make an API request

const response = await fetch(
  "https://openrouter.ai/api/v1/chat/completions",
  {
    method: "POST",
    headers: {
      Authorization: `Bearer ${process.env.OPENROUTER_API_KEY}`,
      "Content-Type": "application/json",
    },
    body: JSON.stringify({
      "model": "qwen/qwen3-235b-a22b-thinking-2507",
      "messages": [
        {
          "role": "user",
          "content": "Explain the trade-offs in this design."
        }
      ]
    }),
  },
);
if (!response.ok) throw new Error(`HTTP ${response.status}`);
const data = await response.json();
console.log(data.choices[0].message.content);

08 / PROVENANCE & HISTORY

Every number has a story.

Follow the source, check the observation date, and distinguish published results from model specifications.

The source register

Direct links to the underlying records

OpenRouter model catalog

Architecture, modalities, parameters, and starting prices

Observed 17 Oct 2026

OpenRouter provider routes

Provider limits, prices, availability, and reported measurements

Observed 17 Oct 2026

Text models ranking

Arena rating · source-reported ranking

Observed 15 Oct 2026

The observation trail

Recorded events, in chronological order

25 Jul 2025Added to the model catalog

Catalog creation date, as recorded by OpenRouter.

17 Oct 2026Model specifications observed

Catalog pricing, limits, and supported parameters captured.

17 Oct 2026Provider routes observed

1 inference routes in this snapshot.

Observation dates describe when data was captured. They do not represent newly run evaluations or continuous performance monitoring.

Published observation ledger

Historical evidence, with source and version preserved.

No published benchmark observations for this selection. Published preference rankings appear separately above.

09 / A FEW MORE DETAILS

Questions worth asking.

Quick answers, drawn from this model’s connected specifications.

How much does qwen3-235b-a22b-thinking-2507 cost?

Catalog starting rates are $0.45 per million input tokens and $3.5 per million output tokens. Compare provider routes and service tiers above for their specific prices.

How much context can it handle?

128,000 tokens in the catalog context window. The catalog lists a maximum of 16,384 completion tokens. Provider limits can differ.

Does it support tools and structured output?

Tool calling: listed as supported. Structured outputs: not listed. See the provider feature matrix for route-specific differences.

How should I interpret these benchmark results?

Read the benchmark name, exact version, metric, source, and model configuration together. Different evaluation settings can change the result. Preference ratings and source-defined composite indices remain separate from task accuracy.

Are these prices and provider measurements live?

This page uses dated source snapshots. Catalog data was observed on 17 Oct 2026. Provider routes were observed on 17 Oct 2026. Missing measurements are not estimated or treated as zero.

KEEP EXPLORING

More from alibaba.

Explore related identities from the same developer, or take them into a side-by-side comparison.

Source-backed evidence. Different tasks. A clearer picture.

About Methodology Benchmark dashboard

Text models

Context window

Input price

Output price

Inference providers

Measured output speed

A clearer view of capability.

The numbers behind the model.

Know what your workload costs.

One model. Different ways to run it.

What goes in. What comes out.

The complete specification sheet.

From exploration to your first request.

Every number has a story.

Questions worth asking.

More from alibaba.

Qwen 3.8 27B

Qwen3.7 Max

Qwen3.7 Plus

Qwen3.8 Flash Next

Qwen3.8 Max

qwen-14b-chat