← Back to model catalog

VENICE ATLAS · MODEL DOSSIER · QWEN 3 235B A22B INSTRUCT 2507

Model overview

Qwen 3 235B A22B Instruct 2507

Built for in-depth research and handling long, complex documents. Ideal for technical work, multimodal input, and high-precision tasks.

qwen3-235b-a22b-instruct-2507 · venice.ai
Text · Venice catalog$0.15 per 1M tokensPrice on Venice AI
ModalityTextOn Venice AI
Primary price$0.15 per 1M tokensper 1M tokens
Context window128KMax tokens
Benchmarks25Benchmark results

Benchmark results

See the tests behind each score. They use the benchmark’s original scoring scale.

Chat3 results
IFEvalifeval · chat
88.7%
ReportedOct 2026
Multi-IFmulti-if · chat
77.5%
ReportedOct 2026
Tau2 Airlinetau2-airline · chat
44.0%
ReportedOct 2026
Communication1 result
Tau2 Retailtau2-retail · communication
71.3%
ReportedOct 2026
Creativity1 result
Creative Writing v3creative-writing-v3 · creativity
87.5%
ReportedOct 2026
Factuality1 result
SimpleQAsimpleqa · factuality
54.3%
ReportedOct 2026
General intelligence6 results
Aider-Polyglotaider-polyglot · general
57.3%
ReportedOct 2026
Arena-Hard v2arena-hard-v2 · general
79.2%
ReportedOct 2026
BFCL-v3bfcl-v3 · general
70.9%
ReportedOct 2026
CSimpleQAcsimpleqa · general
84.3%
ReportedOct 2026
Includeinclude · general
79.5%
ReportedOct 2026
MultiPL-Emultipl-e · general
87.9%
ReportedOct 2026
Language3 results
MMLU-Prommlu-pro · language
83.0%
ReportedOct 2026
MMLU-ProXmmlu-prox · language
79.4%
ReportedOct 2026
MMLU-Reduxmmlu-redux · language
93.1%
ReportedOct 2026
Math4 results
AIME 2025aime-2025 · math
70.3%
ReportedOct 2026
HMMT25hmmt25 · math
55.4%
ReportedOct 2026
LiveBench 20241125livebench-20241125 · math
75.4%
ReportedOct 2026
PolyMATHpolymath · math
50.2%
ReportedOct 2026
Reasoning5 results
ARC-AGIarc-agi · reasoning
41.8%
ReportedOct 2026
GPQAgpqa · reasoning
77.5%
ReportedOct 2026
LiveCodeBench v6livecodebench-v6 · reasoning
51.8%
ReportedOct 2026
SuperGPQAsupergpqa · reasoning
62.6%
ReportedOct 2026
ZebraLogiczebralogic · reasoning
95.0%
ReportedOct 2026
Writing1 result
WritingBenchwritingbench · writing
85.2%
ReportedOct 2026

Learn what each benchmark measures ↗

Related models

More from qwen