Native vision-language MoE model with hybrid linear attention and sparse mixture-of-experts, improved agentic coding, math/code reasoning, and spatial intelligence.
OffRail routes requests to the best providers that are able to handle your prompt size and parameters.
Native vision-language MoE model with hybrid linear attention and sparse mixture-of-experts, improved agentic coding, math/code reasoning, and spatial intelligence. You can access it through OffRail's OpenAI-compatible API with automatic provider routing, fallback, and cost analytics.
Pricing for Qwen3.6 35B A3B on OffRail starts at $0.25 per million input tokens and $1.48 per million output tokens, depending on the provider. The pricing table above always reflects the current per-provider rates.
Qwen3.6 35B A3B supports a context window of up to 262,144 tokens on its largest provider deployment.
Qwen3.6 35B A3B is served by Alibaba Cloud, NovitaAI through OffRail. Requests are automatically routed to the best available provider, with fallback when a provider has issues.
Qwen3.6 35B A3B supports tool (function) calling and soft JSON output, but NOT strict JSON output schema (its upstream provider does not enforce it).
Qwen3.6 35B A3B was released on April 29, 2026.
Qwen Image 3.0
4,500 context$0.00 in / $0.00 out per 1M
Qwen Image 3.0 Pro
4,500 context$0.00 in / $0.00 out per 1M
Qwen3.8 Max
1,000,000 context$1.81 in / $5.45 out per 1M
Qwen3.7 Flash
1,000,000 context$0.03 in / $0.13 out per 1M
Qwen-Audio-3.0-TTS Plus
20,000 context$0.00 in / $0.00 out per 1M
Qwen-Audio-3.0-TTS Flash
20,000 context$0.00 in / $0.00 out per 1M