Alibaba Cloud / Qwenproprietary
Released: 2026-08-01Qwen 3.8 Max Prime
OpenRouter ranked #15 foundation model processing 1.65T tokens weekly. Deployed by alibaba.
Intelligence Index
78.0/ 100
Calibrated multi-domain compositeThroughput Speed
100tok/sec
Output streaming throughputTime To First Token
280ms
Initial chunk server latencyLiveBench Contam-Free
—%
September 2026 suite scoreSourced Benchmark Evaluations (60)
Verified performance across authoritative benchmarks with provenance tracking.| Benchmark | Status | Score | Trust | Date | Source Type | Provenance |
|---|---|---|---|---|---|---|
| ARC-AGI-3 | Active | 50.1% | 93 | 2026-09-05 | independent | Source ↗ |
| ARC-AGI-2 | Nearing Saturation | 60.1% | 75 | 2026-08-29 | independent | Source ↗ |
| FrontierMath | Active | 20.6% | 98 | 2026-08-26 | independent | Source ↗ |
| PutnamBench | Active | 38.2% | 90 | 2026-08-25 | independent | Source ↗ |
| Omni-MATH | Deprecated | 54.4% | 10 | 2026-08-23 | independent | Source ↗ |
| SWE-Lancer | Deprecated | 52.7% | 20 | 2026-08-23 | independent | Source ↗ |
| FrontierCode | Active | 44.9% | 98 | 2026-08-21 | independent | Source ↗ |
| Soohak | Active | 55.9% | 98 | 2026-08-21 | independent | Source ↗ |
| StructFlowBench | Nearing Saturation | 61% | 75 | 2026-08-19 | independent | Source ↗ |
| Humanity's Last Exam | Active | 33.9% | 93 | 2026-08-19 | independent | Source ↗ |
| AIME | Nearing Saturation | 69.8% | 75 | 2026-08-19 | independent | Source ↗ |
| OlympiadBench | Active | 60.3% | 80 | 2026-08-18 | independent | Source ↗ |
| Terminal-Bench | Active | 54.3% | 80 | 2026-08-17 | independent | Source ↗ |
| FireBench | Active | 60.3% | 95 | 2026-08-17 | independent | Source ↗ |
| MathArena | Active | 58.8% | 99 | 2026-08-17 | independent | Source ↗ |
| SimpleQA | Active | 40.6% | 90 | 2026-08-16 | independent | Source ↗ |
| OpenAI-MRCR | Active | 65.5% | 80 | 2026-08-16 | independent | Source ↗ |
| CodeContests | Deprecated | 50.4% | 10 | 2026-08-16 | independent | Source ↗ |
| SWE-bench Pro | Active | 48.1% | 90 | 2026-08-15 | independent | Source ↗ |
| ComplexBench | Nearing Saturation | 58.8% | 80 | 2026-08-15 | independent | Source ↗ |
| LongBench-Pro | Active | 64% | 80 | 2026-08-15 | independent | Source ↗ |
| MultiPL-E | Deprecated | 71.3% | 10 | 2026-08-14 | independent | Source ↗ |
| ZeroSCROLLS | Nearing Saturation | 62.4% | 78 | 2026-08-14 | independent | Source ↗ |
| InfiniteBench | Nearing Saturation | 66.3% | 65 | 2026-08-13 | independent | Source ↗ |
| LiveCodeBench | Saturated | 63.6% | 39 | 2026-08-13 | independent | Source ↗ |
| AgentIF | Nearing Saturation | 48.8% | 80 | 2026-08-13 | independent | Source ↗ |
| MATH | Saturated | 72% | 25 | 2026-08-13 | independent | Source ↗ |
| SCROLLS | Nearing Saturation | 62.4% | 65 | 2026-08-12 | independent | Source ↗ |
| Mercury | Deprecated | 68.2% | 10 | 2026-08-12 | independent | Source ↗ |
| MMLU-Pro | Nearing Saturation | 71.8% | 75 | 2026-08-12 | independent | Source ↗ |
| TabMWP | Saturated | 67.6% | 25 | 2026-08-11 | independent | Source ↗ |
| CyberSecEval | Deprecated | 69.8% | 10 | 2026-08-11 | independent | Source ↗ |
| RULER | Nearing Saturation | 71.8% | 84 | 2026-08-11 | independent | Source ↗ |
| SWE-bench Verified | Saturated | 60.5% | 25 | 2026-08-10 | independent | Source ↗ |
| QuALITY | Nearing Saturation | 68.6% | 78 | 2026-08-10 | independent | Source ↗ |
| MMLU-Redux | Saturated | 76.4% | 25 | 2026-08-10 | independent | Source ↗ |
| SWE-bench | Deprecated | 52.7% | 10 | 2026-08-09 | independent | Source ↗ |
| MGSM | Nearing Saturation | 69.8% | 65 | 2026-08-09 | independent | Source ↗ |
| LongBench | Nearing Saturation | 67.1% | 65 | 2026-08-09 | independent | Source ↗ |
| QASPER | Nearing Saturation | 64.7% | 65 | 2026-08-08 | independent | Source ↗ |
| TruthfulQA | Saturated | 68.6% | 25 | 2026-08-08 | independent | Source ↗ |
| AGIEval | Saturated | 66.2% | 25 | 2026-08-08 | independent | Source ↗ |
| SuperGLUE | Saturated | 78.5% | 25 | 2026-08-07 | independent | Source ↗ |
| GSM-Hard | Deprecated | 70.6% | 10 | 2026-08-07 | independent | Source ↗ |
| MT-Bench | Nearing Saturation | 7.4% | 75 | 2026-08-06 | independent | Source ↗ |
| BIG-Bench Hard | Saturated | 75.5% | 25 | 2026-08-06 | independent | Source ↗ |
| CodeXGLUE | Deprecated | 72.9% | 23 | 2026-08-06 | independent | Source ↗ |
| WinoGrande | Saturated | 80.1% | 28 | 2026-08-05 | independent | Source ↗ |
| MBPP+ | Nearing Saturation | 79.1% | 65 | 2026-08-05 | independent | Source ↗ |
| IFEval | Deprecated | 70.3% | 25 | 2026-08-05 | independent | Source ↗ |
| GPQA | Nearing Saturation | 68.6% | 78 | 2026-08-05 | independent | Source ↗ |
| HumanEval+ | Saturated | 83.7% | 25 | 2026-08-04 | independent | Source ↗ |
| Needle-in-a-Haystack | Saturated | 95.6% | 44 | 2026-08-04 | independent | Source ↗ |
| CommonsenseQA | Saturated | 80.9% | 28 | 2026-08-04 | independent | Source ↗ |
| GSM8K | Saturated | 79.4% | 25 | 2026-08-03 | independent | Source ↗ |
| GLUE | Saturated | 83.2% | 28 | 2026-08-03 | independent | Source ↗ |
| MBPP | Saturated | 82.2% | 25 | 2026-08-03 | independent | Source ↗ |
| HumanEval | Saturated | 88.4% | 25 | 2026-08-02 | independent | Source ↗ |
| MMLU | Saturated | 79.6% | 25 | 2026-08-02 | independent | Source ↗ |
| ARC-AGI | Saturated | 72.4% | 25 | 2026-08-02 | independent | Source ↗ |
OpenAI Compatible API ExecutionAPI Ready
from openai import OpenAI
client = OpenAI()
response = client.chat.completions.create(
model="qwen3.8-max-prime",
messages=[{"role": "user", "content": "Evaluate multi-step logic problem"}],
temperature=0.2
)
print(response.choices[0].message.content)Evaluation & Sourcing Notes
Scores listed for Qwen 3.8 Max Prime represent verified evaluations extracted from official research papers, independent evaluation suites (HELM, LMSYS, OpenCompass, LiveBench), and verified audit reports.
All benchmarks marked as saturated or deprecated reflect historical performance where the benchmark no longer provides active discriminative power.