Qwen3.8 Max

Tool callingReasoningImage inputVideo inputFile input

2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows

Context: 1M
Released: 2mo ago

Sources:models.dev

Benchmarks

DeepSWE is the primary coding benchmark where an exact model evaluation exists. Other sourced evaluations appear below; missing scores are not zero.

Not evaluated in DeepSWE 1.1 as of 2026-09-22. Results are only shown for exact model versions in the official dataset. Official leaderboard.

Other published benchmarks

Evaluation	Reported score and source
FrontierMath-Tier-4-v2-Private	46.3% Epoch AI · qwen3.8-max_xhigh Reasoning effort: xhigh Epoch AI evaluation
FrontierMath-Tiers-1-3-v2-Private	74.7% Epoch AI · qwen3.8-max_xhigh Reasoning effort: xhigh Epoch AI evaluation
GPQA diamond	92.7% Epoch AI · qwen3.8-max_xhigh Reasoning effort: xhigh Epoch AI evaluation
OTIS Mock AIME 2024-2025	99.4% Epoch AI · qwen3.8-max_xhigh Reasoning effort: xhigh Epoch AI evaluation
SimpleQA Verified	45.8% Epoch AI · qwen3.8-max_xhigh Reasoning effort: xhigh Epoch AI evaluation

Benchmark compilation: Epoch AI, Capabilities & benchmarking (CC BY attribution). Scores are converted to percentages and model names matched to the Vertical Tabs catalog. Individual evaluations are credited to their original sources. External results may carry separate licences and evaluation configurations; do not interpret scores from different harnesses as directly comparable.

Specifications

Developer: Alibaba
Model family: Qwen
Type: Language model
Predecessor: —
Successor: —
Released: August 3, 2026
Model ID: alibaba/qwen3.8-max
Openness: Closed
License: —
Input modalities: text, image, video, file
Output modalities: text
Modality: text+image+video+file->text
Max output tokens: 131K
Hugging Face: —
Source freshness: Metadata synced Oct 8, 2026, 10:17 PM UTC

Additional source evidence

Source model: alibaba/qwen3.8-max
Source family: qwen
Last updated: 2026-08-03
Model context: 1M
Model output limit: 131K
Open weights: No
File attachments: Yes
Reasoning: Yes
Tool calling: Yes
Temperature control: Yes

Published benchmarks

Benchmark	Score	Evaluation details	Evidence
Terminal-Bench	86.6 · avg@10	Claude Code · 5h timeout · 2.1	Source
SWE-Bench Pro	67.7 · resolved	Claude Code · Qwen refined and corrected task set	Source
DeepSWE	56.6 · score	Claude Code · 1.1	Source
NL2Repo	55.9 · score	Claude Code	Source
FrontierSWE	73.5 · dominance score	2026-08-03	Source
MLS-Bench-Lite	41 · score	Claude Code · 5h timeout	Source
PaperBench	93 · score	BasicAgent · Code-Dev; 3 runs; 12h timeout; Opus 4.6 judge	Source
AndroidBench	75.1 · avg@3	95 public tasks	Source
QwenSWEBench	80.7 · avg@3	Claude Code · internal	Source
QwenQoderBench	58.4 · avg@5	Claude Code · internal	Source
QwenReactBench	1724 · Elo	internal	Source
QwenSVGBench	1713 · Elo	internal	Source
CoWorkBench	74.8 · score	internal	Source
WorkSpaceBench	67.7 · score	—	Source
JobBench	53.4 · score	—	Source
SkillsBench	70.2 · avg@3	OpenCode · 87 tasks · 1.1	Source
Agents' Last Exam	27 · pass@1	—	Source
Agents' Last Exam	52.4 · score	—	Source
AutomationBench	27.3 · pass@1	600 public tasks	Source
Toolathlon-Verified	72.5 · pass@1	—	Source
WideSearch	81.9 · average item F1	Qwen-Agent	Source
Humanity's Last Exam	56.2 · score	with tools	Source
Humanity's Last Exam	43.6 · score	without tools	Source
GPQA Diamond	92.6 · score	—	Source
IFBench	82.8 · score	—	Source
$OneMillion-Bench	52.5 · expert score	gemini-3.1-pro-preview judge	Source
HealthBench	60.2 · score	—	Source
PLawBench	73.2 · score	gemini-3.1-pro-preview judge	Source
PRBench	57.6 · score	legal	Source
PRBench	58.3 · score	finance	Source
MRCR	92.9 · score	256K; 8-needle · 2	Source
LongBench	66.3 · score	2	Source

Complete source record (JSON)

{
  "id": "alibaba/qwen3.8-max",
  "name": "Qwen3.8 Max",
  "description": "2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows",
  "family": "qwen",
  "last_updated": "2026-08-03",
  "attachment": true,
  "reasoning": true,
  "tool_call": true,
  "temperature": true,
  "modalities": {
    "input": [
      "text",
      "image",
      "video",
      "pdf"
    ],
    "output": [
      "text"
    ]
  },
  "limit": {
    "context": 1000000,
    "output": 131072
  },
  "benchmarks": [
    {
      "name": "Terminal-Bench",
      "score": 86.6,
      "source": "https://huggingface.co/Qwen/Qwen3.8-2.4T-A95B",
      "metric": "avg@10",
      "harness": "Claude Code",
      "variant": "5h timeout",
      "version": "2.1"
    },
    {
      "name": "SWE-Bench Pro",
      "score": 67.7,
      "source": "https://huggingface.co/Qwen/Qwen3.8-2.4T-A95B",
      "metric": "resolved",
      "harness": "Claude Code",
      "dataset": "Qwen refined and corrected task set"
    },
    {
      "name": "DeepSWE",
      "score": 56.6,
      "source": "https://huggingface.co/Qwen/Qwen3.8-2.4T-A95B",
      "metric": "score",
      "harness": "Claude Code",
      "version": "1.1"
    },
    {
      "name": "NL2Repo",
      "score": 55.9,
      "source": "https://huggingface.co/Qwen/Qwen3.8-2.4T-A95B",
      "metric": "score",
      "harness": "Claude Code"
    },
    {
      "name": "FrontierSWE",
      "score": 73.5,
      "source": "https://huggingface.co/Qwen/Qwen3.8-2.4T-A95B",
      "metric": "dominance score",
      "date": "2026-08-03"
    },
    {
      "name": "MLS-Bench-Lite",
      "score": 41,
      "source": "https://huggingface.co/Qwen/Qwen3.8-2.4T-A95B",
      "metric": "score",
      "harness": "Claude Code",
      "variant": "5h timeout"
    },
    {
      "name": "PaperBench",
      "score": 93,
      "source": "https://huggingface.co/Qwen/Qwen3.8-2.4T-A95B",
      "metric": "score",
      "harness": "BasicAgent",
      "variant": "Code-Dev; 3 runs; 12h timeout; Opus 4.6 judge"
    },
    {
      "name": "AndroidBench",
      "score": 75.1,
      "source": "https://huggingface.co/Qwen/Qwen3.8-2.4T-A95B",
      "metric": "avg@3",
      "dataset": "95 public tasks"
    },
    {
      "name": "QwenSWEBench",
      "score": 80.7,
      "source": "https://huggingface.co/Qwen/Qwen3.8-2.4T-A95B",
      "metric": "avg@3",
      "harness": "Claude Code",
      "dataset": "internal"
    },
    {
      "name": "QwenQoderBench",
      "score": 58.4,
      "source": "https://huggingface.co/Qwen/Qwen3.8-2.4T-A95B",
      "metric": "avg@5",
      "harness": "Claude Code",
      "dataset": "internal"
    },
    {
      "name": "QwenReactBench",
      "score": 1724,
      "source": "https://huggingface.co/Qwen/Qwen3.8-2.4T-A95B",
      "metric": "Elo",
      "dataset": "internal"
    },
    {
      "name": "QwenSVGBench",
      "score": 1713,
      "source": "https://huggingface.co/Qwen/Qwen3.8-2.4T-A95B",
      "metric": "Elo",
      "dataset": "internal"
    },
    {
      "name": "CoWorkBench",
      "score": 74.8,
      "source": "https://huggingface.co/Qwen/Qwen3.8-2.4T-A95B",
      "metric": "score",
      "dataset": "internal"
    },
    {
      "name": "WorkSpaceBench",
      "score": 67.7,
      "source": "https://huggingface.co/Qwen/Qwen3.8-2.4T-A95B",
      "metric": "score"
    },
    {
      "name": "JobBench",
      "score": 53.4,
      "source": "https://huggingface.co/Qwen/Qwen3.8-2.4T-A95B",
      "metric": "score"
    },
    {
      "name": "SkillsBench",
      "score": 70.2,
      "source": "https://huggingface.co/Qwen/Qwen3.8-2.4T-A95B",
      "metric": "avg@3",
      "harness": "OpenCode",
      "dataset": "87 tasks",
      "version": "1.1"
    },
    {
      "name": "Agents' Last Exam",
      "score": 27,
      "source": "https://huggingface.co/Qwen/Qwen3.8-2.4T-A95B",
      "metric": "pass@1"
    },
    {
      "name": "Agents' Last Exam",
      "score": 52.4,
      "source": "https://huggingface.co/Qwen/Qwen3.8-2.4T-A95B",
      "metric": "score"
    },
    {
      "name": "AutomationBench",
      "score": 27.3,
      "source": "https://huggingface.co/Qwen/Qwen3.8-2.4T-A95B",
      "metric": "pass@1",
      "dataset": "600 public tasks"
    },
    {
      "name": "Toolathlon-Verified",
      "score": 72.5,
      "source": "https://huggingface.co/Qwen/Qwen3.8-2.4T-A95B",
      "metric": "pass@1"
    },
    {
      "name": "WideSearch",
      "score": 81.9,
      "source": "https://huggingface.co/Qwen/Qwen3.8-2.4T-A95B",
      "metric": "average item F1",
      "harness": "Qwen-Agent"
    },
    {
      "name": "Humanity's Last Exam",
      "score": 56.2,
      "source": "https://huggingface.co/Qwen/Qwen3.8-2.4T-A95B",
      "metric": "score",
      "variant": "with tools"
    },
    {
      "name": "Humanity's Last Exam",
      "score": 43.6,
      "source": "https://huggingface.co/Qwen/Qwen3.8-2.4T-A95B",
      "metric": "score",
      "variant": "without tools"
    },
    {
      "name": "GPQA Diamond",
      "score": 92.6,
      "source": "https://huggingface.co/Qwen/Qwen3.8-2.4T-A95B",
      "metric": "score"
    },
    {
      "name": "IFBench",
      "score": 82.8,
      "source": "https://huggingface.co/Qwen/Qwen3.8-2.4T-A95B",
      "metric": "score"
    },
    {
      "name": "$OneMillion-Bench",
      "score": 52.5,
      "source": "https://huggingface.co/Qwen/Qwen3.8-2.4T-A95B",
      "metric": "expert score",
      "variant": "gemini-3.1-pro-preview judge"
    },
    {
      "name": "HealthBench",
      "score": 60.2,
      "source": "https://huggingface.co/Qwen/Qwen3.8-2.4T-A95B",
      "metric": "score"
    },
    {
      "name": "PLawBench",
      "score": 73.2,
      "source": "https://huggingface.co/Qwen/Qwen3.8-2.4T-A95B",
      "metric": "score",
      "variant": "gemini-3.1-pro-preview judge"
    },
    {
      "name": "PRBench",
      "score": 57.6,
      "source": "https://huggingface.co/Qwen/Qwen3.8-2.4T-A95B",
      "metric": "score",
      "variant": "legal"
    },
    {
      "name": "PRBench",
      "score": 58.3,
      "source": "https://huggingface.co/Qwen/Qwen3.8-2.4T-A95B",
      "metric": "score",
      "variant": "finance"
    },
    {
      "name": "MRCR",
      "score": 92.9,
      "source": "https://huggingface.co/Qwen/Qwen3.8-2.4T-A95B",
      "metric": "score",
      "dataset": "256K; 8-needle",
      "version": "2"
    },
    {
      "name": "LongBench",
      "score": 66.3,
      "source": "https://huggingface.co/Qwen/Qwen3.8-2.4T-A95B",
      "metric": "score",
      "version": "2"
    }
  ],
  "release_date": "2026-08-03",
  "open_weights": false
}

Infrastructure

Provider availability

All providers

Provider endpoint data will appear after this model's first OpenRouter endpoint sync.