Last refreshed 2026-10-19. Next refresh: weekly.
Why use Qwen3-235B-A22B-Instruct-2507 on Together AI?
Together AI offers Qwen3-235B-A22B-Instruct-2507 with pay-as-you-go pricing at $0.20/1M input tokens. Together AI is a platform for running open-source and proprietary LLMs with fast serverless and dedicated endpoints at competitive inference pricing.
Compare Qwen3-235B-A22B-Instruct-2507 across 6 providers to find the best fit for your use caseInput / 1M
$0.20
Output / 1M
$0.60
Cache
Not sourced
Batch
Not sourced
Setup recipe
Python + curlInstall
pip install togetherAuth
export TOGETHER_API_KEY=...Call
from together import Together
client = Together() # reads TOGETHER_API_KEY from env
response = client.chat.completions.create(
model="Qwen/Qwen3-235B-A22B-Instruct-2507-tput",Model ID
Qwen/Qwen3-235B-A22B-Instruct-2507-tputRequest example
from together import Together
client = Together() # reads TOGETHER_API_KEY from env
response = client.chat.completions.create(
model="Qwen/Qwen3-235B-A22B-Instruct-2507-tput",
messages=[{"role": "user", "content": "Hello"}]
)
print(response.choices[0].message.content)Gotchas
- Use provider model ID "Qwen/Qwen3-235B-A22B-Instruct-2507-tput", not the LLMReference slug "qwen3-235b-a22b-instruct-2507".
- Together uses "organization/model-name" format, e.g. "meta-llama/Llama-4-Scout-17B-16E-Instruct" or "Qwen/QwQ-32B". See the Together model catalog for the exact ID.
- The examples expect TOGETHER_API_KEY; rename it only if your application config maps the new variable.
Compare Qwen3-235B-A22B-Instruct-2507 Across Providers
| Provider | Input (per 1M) | Output (per 1M) |
|---|---|---|
| AWS Bedrock | $0.22 | $0.88 |
| Novita AI | $0.09 | $0.58 |
| OpenRouter | $0.09 | $0.55 |
| GCP Vertex AI | $0.22 | $0.88 |
| Together AI | $0.20 | $0.60 |
Pricing
| Type | Price (per 1M) |
|---|---|
| Input tokens | $0.20 |
| Output tokens | $0.60 |
Capabilities
No model capability flags are currently sourced.
About Qwen3-235B-A22B-Instruct-2507
Qwen3-235B-A22B-Instruct-2507 is Alibaba Qwen's July 2025 non-thinking update of Qwen3-235B-A22B, a mixture-of-experts model with 22B active of 235B total parameters. Qwen dropped the hybrid thinking mode and trained separate Instruct and Thinking models; this is the Instruct release, with Apache-2.0 open weights on Hugging Face.
Get Started
Model Specs
Released2025-07-21
Parameters235B-A22B
ArchitectureMixture of Experts