All AI models · LLM benchmarks · Methodology

Alibaba

Qwen3.6 35B A3B FP8

Open multimodal Qwen MoE for local agents that need vision, audio, and code

Model facts

Context window
262K tokens
Maximum output
66K tokens
Cheapest paid input
$0.17 per 1M tokens
Cheapest paid output
$1.1 per 1M tokens
Open weights
yes
Providers
4

Capabilities and modalities

reasoning, tool calling, structured output, attachments, temperature control, open weights, text, image, audio, video.

Local hardware estimate

At 4K context: Q4_K_M 20.2 GiB, Q8_0 36.9 GiB, F16 68.3 GiB working memory.

Parameters
36.0 billion
Architecture
qwen3_5moe
Model license
apache-2.0

Planning estimate adapted from the MIT-licensed whichllm estimator using metadata from Hugging Face; not a vendor minimum requirement.

Observed price history

  • 2026-08-24: $0.17 input / $1.1 output per 1M tokens

API providers

  • Hetzner — model id Qwen/Qwen3.6-35B-A3B-FP8; free input / free output per 1M tokens; provider documentation
  • InferX — model id Qwen3.6-35B-A3B-FP8; free input / free output per 1M tokens; provider documentation
  • NEAR AI Cloud — model id Qwen/Qwen3.6-35B-A3B-FP8; $0.17 input / $1.1 output per 1M tokens; provider documentation
  • Blue Claw — model id Qwen/Qwen3.6-35B-A3B-FP8; not published input / not published output per 1M tokens; provider documentation