BenchmarksAlibabaToday
Qwen3 Max Thinking Benchmark Update
Quality: 39.8/100 | Price: $2.4/M tokens | Output: 42.141 tok/s | HumanEval: 0.431%
View sourceQwen
Qwen3-Max-Thinking is the flagship reasoning model in the Qwen3 series, designed for high-stakes cognitive tasks that require deep, multi-step reasoning. By significantly scaling model capacity and reinforcement learning compute, it...
Running this yourself: can likely run on your own machine.
Qwen3-Max-Thinking is the flagship reasoning model in the Qwen3 series, designed for high-stakes cognitive tasks that require deep, multi-step reasoning.
42.2
Quality Score
---
Arena ELO
Unknown
Parameters
262K
Context
Sign in to join the discussion
0
Downloads
0
Likes
Feb 2026
Released
Benchmarks
19
Open Source
1
Recent launch, pricing, benchmark, and API signals linked to this model or its provider.
Quality: 39.8/100 | Price: $2.4/M tokens | Output: 42.141 tok/s | HumanEval: 0.431%
View sourceQuality: 39.8/100 | Price: $2.4/M tokens | Output: 42.141 tok/s | HumanEval: 0.431%
Quality: 39.8/100 | Price: $2.4/M tokens | Output: 45.366 tok/s | HumanEval: 0.431%
Quality: 39.8/100 | Price: $2.4/M tokens | Output: 46.703 tok/s | HumanEval: 0.431%
View sourceQuality: 39.8/100 | Price: $2.4/M tokens | Output: 47.731 tok/s | HumanEval: 0.431%
View sourceQuality: 39.8/100 | Price: $2.4/M tokens | Output: 45.644 tok/s | HumanEval: 0.431%
View sourceQuality: 39.8/100 | Price: $2.4/M tokens | Output: 45.366 tok/s | HumanEval: 0.431%
Quality: 39.8/100 | Price: $2.4/M tokens | Output: 46.703 tok/s | HumanEval: 0.431%
Quality: 39.8/100 | Price: $2.4/M tokens | Output: 47.731 tok/s | HumanEval: 0.431%
Quality: 39.8/100 | Price: $2.4/M tokens | Output: 45.644 tok/s | HumanEval: 0.431%
Quality: 39.8/100 | Price: $2.4/M tokens | Output: 48.994 tok/s | HumanEval: 0.431%
Quality: 39.8/100 | Price: $2.4/M tokens | Output: 49.38 tok/s | HumanEval: 0.431%
Quality: 39.9/100 | Price: $2.4/M tokens | Output: 48.825 tok/s | HumanEval: 0.431%
Quality: 39.9/100 | Price: $2.4/M tokens | Output: 46.422 tok/s | HumanEval: 0.431%
Quality: 39.9/100 | Price: $2.4/M tokens | Output: 40.787 tok/s | HumanEval: 0.431%
Quality: 39.9/100 | Price: $2.4/M tokens | Output: 44.727 tok/s | HumanEval: 0.431%
Quality: 39.9/100 | Price: $2.4/M tokens | Output: 46.81 tok/s | HumanEval: 0.431%
Quality: 39.9/100 | Price: $2.4/M tokens | Output: 43.876 tok/s | HumanEval: 0.431%
Quality: 39.9/100 | Price: $2.4/M tokens | Output: 35.36 tok/s | HumanEval: 0.431%
Quality: 39.9/100 | Price: $2.4/M tokens | Output: 35.791 tok/s | HumanEval: 0.431%
Quality: 39.9/100 | Price: $2.4/M tokens | Output: 35.171 tok/s | HumanEval: 0.431%
Quality: 39.9/100 | Price: $2.4/M tokens | Output: 35.187 tok/s | HumanEval: 0.431%
Quality: 39.9/100 | Price: $2.4/M tokens | Output: 36.11 tok/s | HumanEval: 0.431%
Quality: 39.9/100 | Price: $2.4/M tokens | Output: 36.02 tok/s | HumanEval: 0.431%
Qwen3 Max Thinking is now available through local Ollama runtime. 40K context window listed. Qwen3 is the latest generation of large language models in Qwen series, offering a comprehensive suite of dense and mixture-of-experts (MoE) models.