Qwen: Qwen3.5-35B-A3B

Image input Video input Text input Text output
Author's Description

The Qwen3.5 Series 35B-A3B is a native vision-language model designed with a hybrid architecture that integrates linear attention mechanisms and a sparse mixture-of-experts model, achieving higher inference efficiency. Its overall...

Key Specifications
Cost
$$$$$
Context
262K
Parameters
35B
Released
Feb 25, 2026
Speed
Ability
Reliability
Supported Parameters

This model supports the following parameters:

Top Logprobs Response Format Top P Seed Temperature Logprobs Reasoning Max Tokens Tool Choice Structured Outputs Tools Include Reasoning Presence Penalty
Features

This model supports the following features:

Reasoning Structured Outputs Tools Response Format
Performance Summary

The Qwen3.5-35B-A3B model, a native vision-language model with a hybrid architecture, demonstrates exceptional speed and cost efficiency. It consistently ranks among the fastest models, achieving an Infinityth percentile in speed across 8 benchmarks, and offers highly competitive pricing, also at the Infinityth percentile across 4 benchmarks. The model exhibits strong reliability with an 88% success rate across 8 benchmarks, indicating consistent operational performance. In terms of specific benchmarks, Qwen3.5-35B-A3B shows a notable strength in Hallucinations (Baseline) with 98.0% accuracy, placing it in the 63rd percentile, and in Email Classification (Baseline) with 99.0% accuracy (73rd percentile). Its Instruction Following (Baseline) capability is also strong at 74.7% accuracy (80th percentile). However, the model currently shows significant weaknesses in General Knowledge, Coding, Reasoning, Ethics, and Mathematics, where it recorded 0.0% accuracy across all these categories. This suggests a specialized performance profile, excelling in specific language understanding and classification tasks while requiring further development in complex reasoning and domain-specific knowledge areas.

Model Pricing

Current Pricing

Feature Price (per 1M tokens)
Prompt $0.25
Completion $1

Price History

Available Endpoints
Provider Endpoint Name Context Length Pricing (Input) Pricing (Output)
Alibaba
Alibaba | qwen/qwen3.5-35b-a3b-20260224 262K $0.163 / 1M tokens $1.3 / 1M tokens
Parasail
Parasail | qwen/qwen3.5-35b-a3b-20260224 262K $0.25 / 1M tokens $1 / 1M tokens
Venice
Venice | qwen/qwen3.5-35b-a3b-20260224 256K $0.313 / 1M tokens $1.25 / 1M tokens
AtlasCloud
AtlasCloud | qwen/qwen3.5-35b-a3b-20260224 262K $0.225 / 1M tokens $1.8 / 1M tokens
Ionstream
Ionstream | qwen/qwen3.5-35b-a3b-20260224 262K $0.163 / 1M tokens $1.3 / 1M tokens
NextBit
NextBit | qwen/qwen3.5-35b-a3b-20260224 262K $0.3 / 1M tokens $1.8 / 1M tokens
AkashML
AkashML | qwen/qwen3.5-35b-a3b-20260224 262K $0.23 / 1M tokens $1.8 / 1M tokens
Benchmark Results
Benchmark Category Reasoning Strategy Free Executions Accuracy Cost Duration
Other Models by qwen