AiCostCompare

Qwen3 235B A22B Instruct 2507

by Alibaba (Qwen) · qwen/qwen3-235b-a22b-2507 · #73 cheapest of 422 paid models

Prices updated Sep 19, 2026, 10:37 PM UTC · refreshed hourly

Qwen3-235B-A22B-Instruct-2507 is a multilingual, instruction-tuned mixture-of-experts language model based on the Qwen3-235B architecture, with 22B active parameters per forward pass. It is optimized for general-purpose text generation, including instruction following,...

Compare vs:Claude Sonnet 5Claude Opus 4.6Claude Haiku 4.5GPT-5.6 Sol

Pricing

Input

$0.087

per 1M tokens

Output

$0.35

per 1M tokens

Blended (3:1)

$0.153

3 input : 1 output

Cache read

$0.018

per 1M cached tokens

Cache write

per 1M tokens

Cache & batch economics

Effective prices for Qwen3 235B A22B Instruct 2507 with prompt caching.

Cache hit rateEffective blended $/1MExample request*vs no cache
0%$0.153$0.009275
50%$0.127$0.005775−38%
90%$0.106$0.002975−68%

*Example: 100K sticky context + 2K new input + 1K output. Batch discount is an approximate 50% off list rates (available for OpenAI / Anthropic / Google — not flagged for this creator). Confirm on the vendor’s pricing page.

What a request costs

WorkloadInput tokensOutput tokensCost / requestCost / 1K requests
Short chat message500300$0.000149$0.1487
Document summary8,0001,000$0.00105$1.05
Codebase question (RAG)30,0002,000$0.003325$3.32
Long-context analysis150,0005,000$0.0149$14.88

Performance

Intelligence Index

12

AA composite quality

Coding Index

Math Index

71.7

Agentic Index

Output speed

0 tok/s

median

Time to first token

0s

median

Time to first answer

0s

after reasoning tokens

Artificial Analysis benchmarks

BenchmarkScore
MMLU-Pro82.8%
GPQA Diamond75.3%
Humanity's Last Exam11.1%
LiveCodeBench52.4%
MATH-50098.0%
AIME71.7%
AIME 202571.7%
IFBench46.1%
AA-LCR (Long Context Reasoning)33.9%
Terminal-Bench Hard15.2%
τ²-Bench (Telecom)33.3%

Design Arena Elo

CategoryElo
Website1,070
UI components980
Data viz1,084
3D1,023
Game dev975
Code categories1,053

Specs

Context window

262K

tokens

Max output

236K

tokens

Input modalities

text

Tokenizer

Qwen3