Alibaba Cloud / Qwen Team model product
Qwen3 Max Thinking
Qwen3 Max Thinking is a reasoning-focused large language model in the Qwen3 family from Alibaba Cloud / Qwen Team.
Updated Oct 23, 2026. Default version: Qwen3 Max Thinking
LLMBoard Score53.5Qwen3 Max Thinking
Coverage100%33 benchmark families
Context window256KTokens
Official input priceN/AOfficial price unavailable
Qwen3 Max Thinking Capability Profile
This profile uses the model's current scored version. Arena ratings and prices are shown separately.
Qwen3 Max Thinking Benchmark Results
Benchmark scores for Qwen3 Max Thinking.
Sort by
Benchmark | Score | Rank | Participants | Percentile | Evidence | Evaluated |
|---|
| BenchmarkC-Eval | Score93.70% | Rank01 | Participants20 | Percentile100.00% | EvidenceC | Evaluated |
| BenchmarkSecCodeBench | Score57.50% | Rank02 | Participants2 | Percentile0.00% | EvidenceC | Evaluated |
| BenchmarkHLE-Verified | Score37.60% | Rank03 | Participants3 | Percentile0.00% | EvidenceC | Evaluated |
| BenchmarkDeepPlanning | Score28.70% | Rank04 | Participants10 | Percentile66.67% | EvidenceC | Evaluated |
| BenchmarkSeal-0 | Score46.90% | Rank05 | Participants7 | Percentile33.33% | EvidenceC | Evaluated |
| BenchmarkVITA-Bench | Score40.90% | Rank06 | Participants11 | Percentile50.00% | EvidenceC | Evaluated |
| BenchmarkHumanity's Last Exam (with tools, text-only) | Score49.80% | Rank07 | Participants10 | Percentile33.33% | EvidenceC | Evaluated |
| BenchmarkIFEval | Score93.40% | Rank07 | Participants70 | Percentile91.30% | EvidenceC | Evaluated |
| BenchmarkSuperGPQA | Score67.30% | Rank07 | Participants37 | Percentile83.33% | EvidenceC | Evaluated |
| BenchmarkBFCL-V4 | Score67.70% | Rank08 | Participants22 | Percentile66.67% | EvidenceC | Evaluated |
| BenchmarkMAXIFE | Score84.00% | Rank08 | Participants12 | Percentile36.36% | EvidenceC | Evaluated |
| BenchmarkMCP-Mark | Score33.50% | Rank09 | Participants9 | Percentile0.00% | EvidenceC | Evaluated |
| BenchmarkMulti-Challenge | Score63.30% | Rank09 | Participants34 | Percentile75.76% | EvidenceC | Evaluated |
| BenchmarkNOVA-63 | Score54.20% | Rank10 | Participants12 | Percentile18.18% | EvidenceC | Evaluated |
| BenchmarkAA-LCR | Score68.70% | Rank11 | Participants27 | Percentile61.54% | EvidenceC | Evaluated |
| BenchmarkBrowseComp-zh | Score60.90% | Rank11 | Participants15 | Percentile28.57% | EvidenceC | Evaluated |
| BenchmarkGlobal PIQA | Score86.00% | Rank11 | Participants15 | Percentile28.57% | EvidenceC | Evaluated |
| BenchmarkHumanity's Last Exam (no tools, text-only) | Score30.20% | Rank11 | Participants14 | Percentile23.08% | EvidenceC | Evaluated |
| BenchmarkIMO-AnswerBench | Score83.90% | Rank11 | Participants22 | Percentile52.38% | EvidenceC | Evaluated |
| BenchmarkWMT24++ | Score77.60% | Rank11 | Participants24 | Percentile56.52% | EvidenceC | Evaluated |
| BenchmarkLongBench v2 | Score60.60% | Rank12 | Participants19 | Percentile38.89% | EvidenceC | Evaluated |
| BenchmarkAIME 2026 | Score93.30% | Rank14 | Participants27 | Percentile50.00% | EvidenceC | Evaluated |
| BenchmarkLiveCodeBench v6 | Score85.90% | Rank14 | Participants67 | Percentile80.30% | EvidenceC | Evaluated |
| BenchmarkWideSearch | Score57.90% | Rank14 | Participants15 | Percentile7.14% | EvidenceC | Evaluated |
| BenchmarkMMLU-Redux | Score92.80% | Rank16 | Participants50 | Percentile69.39% | EvidenceC | Evaluated |
| BenchmarkMMLU-Pro | Score85.70% | Rank17 | Participants142 | Percentile88.65% | EvidenceC | Evaluated |
| BenchmarkIFBench | Score70.90% | Rank29 | Participants46 | Percentile37.78% | EvidenceC | Evaluated |
| BenchmarkMMMLU | Score84.40% | Rank34 | Participants51 | Percentile34.00% | EvidenceC | Evaluated |
| BenchmarkSWE-bench Multilingual | Score66.70% | Rank35 | Participants50 | Percentile30.61% | EvidenceC | Evaluated |
| BenchmarkSWE-Bench Verified | Score75.30% | Rank40 | Participants120 | Percentile67.23% | EvidenceC | Evaluated |
Qwen3 Max Thinking Arena Results
Preference and agent-evaluation results for the default version.
No Arena results
The default version does not have a matching Arena result yet.
Qwen3 Max Thinking Pricing
Official vendor API pricing appears first, followed by individual provider offers.
- Official API
- N/A
- Official provider
- N/A
- Lowest third-party
- From $1.2 input, $6 output per 1M via ZenMux
- Tracked offerings
- 2
Sort by
Provider | Provider model ID | Region | Input / 1M | Output / 1M | Context | Updated |
|---|
| ProviderZenMux | Provider model IDqwen/qwen3-max | Regionglobal | Input / 1M$1.2 | Output / 1M$6 | Context256K | Updated |
| ProviderVercel AI Gateway | Provider model IDalibaba/qwen3-max-thinking | Regionglobal | Input / 1M$1.2 | Output / 1M$6 | Context256K | Updated |
Official prices use only the vendor's configured official Provider and positive standard USD PAYG rates. Third-party offers remain explicitly labeled.
Qwen3 Max Thinking Runtime Performance
Provider-specific output speed and catalog latency for Qwen3 Max Thinking. Runtime does not affect the capability score.
No runtime data
No provider-specific speed or latency record is linked to the default version yet.
Output Speed is generated output tokens received per second. Catalog latency is reported separately from observed provider TTFT.
Qwen3 Max Thinking Specifications
Technical details for the model's default version.
- Version
- Qwen3 Max Thinking
- Released
- Feb 13, 2026
- Knowledge cutoff
- Unknown
- Parameters
- 1T
- Context window
- 256K
- Max output
- 256K
- Inputs
- text
- Outputs
- text
- Open weights
- No
- License
- Proprietary
Qwen3 Max Thinking Versions
Available versions of this model. The score column identifies the version used in the overall ranking.
Sort by
Version | Released | LLMBoard | Parameters | Context | Max output | Open weights | License |
|---|
| VersionQwen3 Max Thinking | Released | LLMBoard53.55 | Parameters1T | Context256K | Max output256K | Open weightsNo | LicenseProprietary |
Models similar to Qwen3 Max Thinking
Recommendations prioritize the same model type and family, then the closest LLMBoard score.
AC
Qwen3.6 35B A3B
Alibaba Cloud / Qwen Team
AC
Qwen3.5 27B
Alibaba Cloud / Qwen Team
AC
Qwen3.5 35B A3B
Alibaba Cloud / Qwen Team
AC
Qwen3.5 122B A10B
Alibaba Cloud / Qwen Team
AC
Qwen3.6 27B
Alibaba Cloud / Qwen Team
AC
Qwen3 235B A22B Thinking
Alibaba Cloud / Qwen Team
What is Qwen3 Max Thinking?
Key information about Qwen3 Max Thinking and its available data.
Qwen3 Max Thinking is a large language model in the Qwen3 family with adaptive tool-use and advanced test-time scaling techniques.
Data as of 2026-10-22.
FAQ
Common questions about Qwen3 Max Thinking.
When was Qwen3 Max Thinking released?
Qwen3 Max Thinking's default version was released on Feb 13, 2026.
How much does Qwen3 Max Thinking cost?
No official standard PAYG price is currently available for Qwen3 Max Thinking. The lowest tracked third-party offer starts at $1.2 input and $6 output via ZenMux.
Who created Qwen3 Max Thinking?
Qwen3 Max Thinking was created by Alibaba Cloud / Qwen Team.
What is the context window for Qwen3 Max Thinking?
The default version has a 256K token context window.
Is Qwen3 Max Thinking open weight?
No. The default version is not marked as having publicly available weights.