qwen
Qwen: Qwen3 235B A22B Thinking 2507
qwen/qwen3-235b-a22b-thinking-2507
Qwen3-235B-A22B-Thinking-2507 is a high-performance, open-weight Mixture-of-Experts (MoE) language model optimized for complex reasoning tasks. It activates 22B of its 235B parameters per forward pass and natively supports up to 262,144...
- Type
- text
- Provider
- qwen
- Model ID
- qwen/qwen3-235b-a22b-thinking-2507
- Capabilities
- tools, reasoning
- Context window
- 131K tokens
- Self-hostable
- Yes — runs on your own GPU
- Input price
- $0.25 / 1M tokens
- Output price
- $2.53 / 1M tokens
Cloud price is billed from prepaid credits when a request fails over to the cloud. Open-weight models run free on GPUs you own — the gateway routes to your nodes first.
What it costs to run
cloud · 1M in + 1M out
$2.78
$0.25 for a million input tokens plus $2.53 for a million output tokens, billed from credits only when a request fails over to the cloud.
your hardware
$0 / token
Served from a workstation or on-prem server you own, there is no per-token fee — the marginal cost is power.
Run it on your own hardware
Estimated GPU memory for Qwen: Qwen3 235B A22B Thinking 2507 (235B parameters), for the weights alone — add room for the context cache.
- Q4_K_M
- ~134.3 GBmulti-GPU or large unified memory
- Q8_0
- ~232.7 GBmulti-GPU or large unified memory
- FP16 / BF16
- ~438.5 GBmulti-GPU or large unified memory
Install the agent on the machine and it dials out to the gateway over a single WebSocket — no port forwarding or inbound firewall rules. Requests for this model go to your node first and fail over to the cloud only when it can't serve.
/// initialize
Call Qwen: Qwen3 235B A22B Thinking 2507 through one OpenAI-compatible endpoint.
Serve it from hardware you own, with cloud failover when your nodes are busy or offline.
no credit card · 2 nodes free · openai-compatible
Frequently asked questions
- What is the context window of Qwen: Qwen3 235B A22B Thinking 2507?
- Qwen: Qwen3 235B A22B Thinking 2507 accepts up to 131,072 tokens (131K) of context per request.
- How much does Qwen: Qwen3 235B A22B Thinking 2507 cost per million tokens?
- Through Wide Area Intelligence, Qwen: Qwen3 235B A22B Thinking 2507 costs $0.25 per 1M input tokens and $2.53 per 1M output tokens when served from the cloud, billed from prepaid credits. A workload of 1M input plus 1M output tokens costs $2.78.
- Can I run Qwen: Qwen3 235B A22B Thinking 2507 on my own hardware?
- Yes — Qwen: Qwen3 235B A22B Thinking 2507 is an open-weight model you can run on a workstation or on-prem server you own. At Q4_K_M it needs roughly 134.3 GB of GPU memory, plus room for the context cache — more than a single 80 GB card, so plan for multiple GPUs or a large unified-memory machine. Wide Area Intelligence routes requests to your own nodes first and fails over to the cloud only when they can't serve.
More from qwen
- Qwen2.5 72B Instructqwen/qwen-2.5-72b-instruct
- Qwen: Qwen Plus 0728qwen/qwen-plus-2025-07-28
- Qwen: Qwen-Plusqwen/qwen-plus
- Qwen: Qwen2.5 7B Instructqwen/qwen-2.5-7b-instruct
- Qwen: Qwen2.5 VL 72B Instructqwen/qwen2.5-vl-72b-instruct
- Qwen: Qwen3 14Bqwen/qwen3-14b
- Qwen: Qwen3 235B A22Bqwen/qwen3-235b-a22b
- Qwen: Qwen3 235B A22B Instruct 2507qwen/qwen3-235b-a22b-2507