Qwen3 Max

qwen/qwen3-max
ToolsJSONReasoning
by Qwen · 2025-09-23

Qwen3 Max — proprietary flagship chat model, 256k context, thinking mode + function calling.

ctx262.1K tokens
Max output65.5K
Inputtext
Outputtext
p50 TTFT1.92 s
INPUT$0.36/ 1M tokens
OUTPUT$1.43/ 1M tokens
p50 TTFT1.92 s7d
p95 TTFT10.00 s7d
TRAFFIC768.3Ktokens / 7d

Qwen3 Max is a Mixture-of-Experts (MoE) language model from Alibaba's Qwen team. It is designed for high-capacity tasks that require extended context and deep reasoning. The model accepts text-only…

What is Qwen3 Max?

Who should use Qwen3 Max?

What is OrcaRouter's role?

What input modalities are supported?

Code samples

Call from any SDK

OpenAI-compatible — keep the SDK you already use

  • OpenAI SDKhttps://api.orcarouter.ai/v1
from openai import OpenAI

client = OpenAI(
    base_url="https://api.orcarouter.ai/v1",
    api_key="$ORCAROUTER_API_KEY",
)

response = client.chat.completions.create(
    model="qwen/qwen3-max",
    messages=[{"role": "user", "content": "Hello"}],
)
print(response.choices[0].message.content)

Supported parameters

  • enable_search
  • enable_thinking
  • include_reasoning
  • logprobs
  • max_tokens
  • n
  • parallel_tool_calls
  • presence_penalty
  • reasoning
  • repetition_penalty
  • response_format
  • seed
  • stop
  • stream
  • stream_options
  • temperature
  • thinking_budget
  • tool_choice
  • tools
  • top_k
  • top_logprobs
  • top_p

Pricing

TierInput / 1M tokensOutput / 1M tokens
32K$0.359$1.434
128K$0.574$2.294
256K$1.004$4.014
Tier selected by input token count of each request

Cost calculator

Tokens / month10MM
Input share70%%
Estimated / month $6.82

Estimate based on list price

Tiered pricing — this estimate uses base-tier rates.

Token & cost estimator

Input tokens: 20Cost per request: $0.000724

Estimate only — actual token counts depend on the provider's tokenizer.

Performance

p50 TTFT
1.92 s
Output speed
86.8 tok/s
p95 TTFT
10.00 s
Error rate
0%

Public benchmarks

26.4
AA Coding
Better than 29% of models compared
#75 of 106
31.4
AA Intelligence
Better than 32% of models compared
#75 of 110
80.7
AA Math
Better than 70% of models compared
#23 of 81
AIME 2025
80.7
GPQA Diamond
76.4
Humanity's Last Exam
11.1
IFBench
44.1
LiveCodeBench
76.7
Long-Context Recall
46.7
MMLU-Pro
84.1
SciCode
38.3
TerminalBench Hard
20.5
τ²-Bench
74.3
Source: artificialanalysis.ai

How it compares

Qwen3 Maxqwen/qwen3-max-previewQwen3.5 397B A17Bqwen/qwen3.5-plus
Input $/M$0.36$0.86$0.17$0.12
Output $/M$1.43$3.44$1.03$0.69
Context262K262K33K1.0M
Quality7/108/108/108/10
Compare side-by-sideCompare side-by-sideCompare side-by-sideCompare side-by-side

FAQ

What is the cost per token for Qwen3 Max on OrcaRouter?
Per-token pricing for Qwen3 Max is published on OrcaRouter's pricing page and varies by usage tier. There is no flat fee; you pay only for input and output tokens. Because Qwen3 Max is a large MoE model, its per-token cost is higher than smaller models but competitive with other top-tier models. Check OrcaRouter's website for current rates.
What is the context window size of Qwen3 Max?
Qwen3 Max supports a context window of 262,144 tokens, meaning it can accept up to that many tokens as input (including system and user messages). The maximum output is 65,536 tokens per request.
What are the main strengths of Qwen3 Max?
Its main strengths are a very large context window (262k tokens), high output limit (65k tokens), strong performance on professional reasoning benchmarks (MMLU-Pro 84.1), and an MoE architecture that balances capability with inference efficiency. It is ideal for long-document analysis and complex multi-step reasoning.
How does Qwen3 Max compare to GPT-4 or Claude 3?
Comparisons depend on the specific variant. Qwen3 Max has a larger context window than most GPT-4 versions (128k for GPT-4 Turbo) and higher output limit than Claude 3 Opus (4k). Benchmark scores on MMLU-Pro are comparable to top-tier models, but actual performance varies by task. OrcaRouter offers multiple models; you can test side-by-side.
Does OrcaRouter train on my data when I use Qwen3 Max?
No. OrcaRouter does not use customer data for training or improving models. All inputs and outputs are processed solely for inference and logged for billing and operational purposes. Data is not shared with third parties beyond the necessary infrastructure. See OrcaRouter's privacy policy for details.
How do I call Qwen3 Max using an OpenAI-compatible API?
Set your base URL to https://api.orcarouter.ai/v1, provide your OrcaRouter API key, and use the model ID 'qwen/qwen3-max'. All OpenAI SDKs and direct HTTP clients work without modification. Example: client = OpenAI(base_url='https://api.orcarouter.ai/v1', api_key='...') then client.chat.completions.create(model='qwen/qwen3-max', messages=[...]).
What are the supported input and output modalities?
Qwen3 Max accepts text-only input and generates text-only output. It does not process images, audio, or video. For multimodal tasks, consider models like Qwen2-VL available on OrcaRouter.
Can I use function calling with Qwen3 Max?
Yes, Qwen3 Max supports the OpenAI-compatible function calling and tool use format. You can define functions in the 'tools' parameter, and the model can request to call them. This works through OrcaRouter's API without any extra configuration.
Is there a rate limit for Qwen3 Max on OrcaRouter?
OrcaRouter applies rate limits to ensure fair usage. These limits are typically based on tokens per minute and requests per minute. Exact limits depend on your plan. Check OrcaRouter's documentation or your dashboard for specific rates.
What are the limitations of Qwen3 Max?
Like all LLMs, it may hallucinate or produce incorrect information, especially on obscure topics. It has a training cutoff (not publicly disclosed), so it cannot access real-time events without context. The large context can lead to 'lost in the middle' effects. It is text-only and not designed for real-time applications without streaming.

Embed this badge

Qwen: Qwen3 Max$0.36/M in1916ms p50via OrcaRouter
HTML <a href="https://www.orcarouter.ai/models/qwen/qwen3-max" target="_blank"> <img src="https://www.orcarouter.ai/embed/qwen/qwen3-max.svg" alt="Qwen: Qwen3 Max on OrcaRouter" /> </a>
Markdown [![Qwen: Qwen3 Max](https://www.orcarouter.ai/embed/qwen/qwen3-max.svg)](https://www.orcarouter.ai/models/qwen/qwen3-max)

Model card as data

GET /api/public/models/qwen/qwen3-maxOpen
Machine-readable:/llms.txt/llms-full.txt