AlibabaQwen3.8 Max
Qwen3.8 Max is an Alibaba proprietary reasoning model, released 3 Aug 2026. Its best configuration (max reasoning effort) ranks #52 of 760 on the BenchLeader Index at 64.2 ±3.4, in the upper half. The ± is the point: 109 other configurations score within that range, so they and this one cannot be told apart on quality alone — price and speed are what separate them. It scores highest in reasoning (72) and lowest in agents & tools (58). At $3.00 per million tokens blended it is pricier than most ranked models. Output speed of 37 tokens per second puts it in the slowest quarter, with a first answer in 58.9 s. It has been measured at 3 reasoning-effort settings; this summary describes the best-scoring one, and the tabs above switch between them. Last measured 9 Oct 2026.
- Blended price
- $3.00/M
- $2.00 in · $6.00 out
- Output speed
- 37 tok/s
- measured by Artificial Analysis
- First answer
- 59 s
- first token 2.70 s
- Context
- 1M
- Full answer
- 73 s
- median, reasoning included
- Cost per run
- $5.41
- one full Intelligence Index run
- Released
- 3 Aug 2026
- Overall index64
- Reasoning72
- Coding65
- Agents & tools58
- Maths65
- Knowledge63
- Human preference68
- Multimodal66
- Long context65
- Composite71
Versions
Alibaba has shipped 2 models under this name. Each is ranked on its own results; a newer version often has fewer results so far, which holds its index nearer the average until more arrive.
| Model | Released | Index | Rank |
|---|---|---|---|
| Qwen3.8 Maxmaxthis page | 3 Aug 2026 | 64.2 | #52 |
| Qwen3.6 Maxmax | 20 Apr 2026 | 62.5 | #77 |
Reasoning-effort configurations
The same model behaves differently depending on how much it is allowed to think. Each row is one setting, scored only on the benchmarks that were run at that setting. “Not stated” collects results from publishers that did not say which setting they used; for a reasoning model that is usually its thinking mode, but we do not assume it. Pick a setting here or at the top of the page to see its price, speed and category scores.
| Effort | Index | Rank | Speed | First answer | Chat reply cost | Categories |
|---|---|---|---|---|---|---|
| xhigh | 57.3 | #170 | 35 tok/s | 2.70 s | $0.0026 | Knowledge 55 · Maths 63 · Reasoning 67 |
| maxbest | 64.2 | #52 | 37 tok/s | 59 s | $0.0026 | Agents & tools 58 · Coding 65 · Composite 71 · Human preference 68 · Knowledge 63 · Long context 65 · Maths 65 · Multimodal 66 · Reasoning 72 |
| not stated | – | – | 35 tok/s | 2.70 s | $0.0026 | Coding 67 |
How its index has moved
27 Sept 2026 to 8 Oct 2026- max
- xhigh
The index is recomputed from scratch every day, so a line moves when a new benchmark result lands, when a publisher revises a score, or when the models it is normalised against change. Early movement usually means the score is still settling.
Benchmark results
One column per reasoning effort. Rank is among every configuration of every model on that benchmark. Hover a score for the run it came from.
Reasoning
| Benchmark | xhigh | max | not stated | Source | Trend |
|---|---|---|---|---|---|
| GPQA Diamond | 92.7%#24 | – | – | Epoch AI Benchmarking Hub | |
| LMArena Hard Prompts | – | 1504#25 | – | LMArena | |
| LiveBench Reasoningnot in index | – | 88.2%#22 | – | LiveBench | |
| GPQA Diamond (AA)not in index | – | 92.8%#28 | – | Artificial Analysis | |
| Humanity's Last Exam (AA)not in index | – | 43.1%#55 | – | Artificial Analysis | |
| GPQA Diamond (Vals)not in index | – | 93.7%#6 | – | Vals AI | |
| CritPt | – | 20.0%#57 | – | Artificial Analysis | |
| MysteryMechanismnot in index | – | 23.9%#14 | – | Vals AI | |
| Chess Puzzlesnot in index | 40.0%#24 | – | – | Epoch AI Benchmarking Hub | |
| Mystery Game Puzzlesnot in index | 38.0%#15 | – | – | Epoch AI Benchmarking Hub | |
| LMCAnot in index | 46.2%#69 | – | – | Epoch AI Benchmarking Hub | |
| DTBenchnot in index | 92.0%#56 | – | – | Epoch AI Benchmarking Hub |
Coding
| Benchmark | xhigh | max | not stated | Source | Trend |
|---|---|---|---|---|---|
| SciCode | – | 53.2%#73 | 52.1%#79 | SciCode | |
| LMArena Coding | – | 1524#29 | – | LMArena | |
| LMArena WebDev | – | 1672#11 | 1674#10 | LMArena | |
| LiveBench Codingnot in index | – | 72.9%#54 | – | LiveBench | |
| SciCode (AA)not in index | – | 53.2%#73 | – | Artificial Analysis | |
| LiveCodeBench | – | 87.8%#10 | – | Vals AI | |
| IOI | – | 68.9%#11 | – | Vals AI | |
| SWE-bench (Vals)not in index | – | 85.6%#17 | – | Vals AI | |
| Code Migrationnot in index | – | 24.0%#41 | – | Vals AI | |
| ProgramBenchnot in index | – | 0.0%#23 | – | Vals AI | |
| Vibe Code Bench 1-100not in index | – | 12.8%#17 | – | Vals AI | |
| Vibe Code Bench v1.1not in index | – | 64.7%#38 | – | Vals AI | |
| DeepSWE v1.1not in index | 57.5%#36 | – | – | Epoch AI Benchmarking Hub | |
| FrontierSWEnot in index | 17.8%#16 | – | – | Epoch AI Benchmarking Hub | |
| Terminal-Bench 4.0 (AA)not in index | – | 38.9%#30 | – | Artificial Analysis | |
| Terminal-Bench 2.1 (AA)not in index | – | 88.8%#9 | – | Artificial Analysis |
Agents & tools
Maths
| Benchmark | xhigh | max | not stated | Source | Trend |
|---|---|---|---|---|---|
| FrontierMath Tiers 1–3 | 74.7%#19 | – | – | Epoch AI Benchmarking Hub | |
| FrontierMath Tier 4 | 46.3%#25 | – | – | Epoch AI Benchmarking Hub | |
| OTIS Mock AIME | 100.0%#1 | – | – | Epoch AI Benchmarking Hub | |
| ProofBench | – | 58.0%#19 | – | Vals AI | |
| LiveBench Mathematicsnot in index | – | 91.3%#29 | – | LiveBench | |
| LMArena Maths | – | 1497#19 | – | LMArena |
Knowledge
| Benchmark | xhigh | max | not stated | Source | Trend |
|---|---|---|---|---|---|
| SimpleQA Verified | 47.3%#32 | – | – | Epoch AI Benchmarking Hub | |
| LiveBench Data Analysisnot in index | – | 78.4%#23 | – | LiveBench | |
| AA-Omniscience | – | 12.0#86 | – | Artificial Analysis | |
| MMLU-Pro | – | 88.6%#18 | – | Vals AI | |
| LegalBench | – | 83.6%#48 | – | Vals AI | |
| CorpFin | – | 65.8%#26 | – | Vals AI | |
| TaxEval | – | 75.5%#17 | – | Vals AI | |
| Excel Modeling Benchmarknot in index | – | 60.1%#32 | – | Vals AI | |
| MedCodenot in index | – | 40.7%#57 | – | Vals AI | |
| MedScribenot in index | – | 85.0%#30 | – | Vals AI | |
| Public Benefits Benchnot in index | – | 67.1%#13 | – | Vals AI | |
| GDP.pdfnot in index | 23.2%#19 | – | – | Epoch AI Benchmarking Hub | |
| AA-Omniscience: accuracynot in index | – | 31.9%#161 | – | Artificial Analysis | |
| AA-Omniscience: non-hallucinationnot in index | – | 71.2%#34 | – | Artificial Analysis |
Instruction following
| Benchmark | xhigh | max | not stated | Source | Trend |
|---|---|---|---|---|---|
| LiveBench Languagenot in index | – | 79.7%#33 | – | LiveBench | |
| LiveBench Instruction Followingnot in index | – | 74.1%#13 | – | LiveBench | |
| LMArena Instruction Followingnot in index | – | 1474#28 | – | LMArena |
Human preference
| Benchmark | xhigh | max | not stated | Source | Trend |
|---|---|---|---|---|---|
| LMArena Text | – | 1483#22 | – | LMArena | |
| LMArena Creative Writingnot in index | – | 1470#16 | – | LMArena | |
| LMArena Multi-turnnot in index | – | 1492#19 | – | LMArena | |
| LMArena Longer Queriesnot in index | – | 1492#25 | – | LMArena |
Multimodal
| Benchmark | xhigh | max | not stated | Source | Trend |
|---|---|---|---|---|---|
| LMArena Vision | – | 1314#9 | – | LMArena | |
| MMMU-Pro | – | 82.8%#29 | – | Artificial Analysis | |
| MMMU-Pro (Vals)not in index | – | 88.0%#12 | – | Vals AI | |
| MortgageTaxnot in index | – | 64.0%#48 | – | Vals AI | |
| SAGEnot in index | – | 51.3%#14 | – | Vals AI |
Long context
| Benchmark | xhigh | max | not stated | Source | Trend |
|---|---|---|---|---|---|
| AA-LCR | – | 80.3%#66 | – | Artificial Analysis | |
| MLCRnot in index | – | 20.0%#19 | – | Artificial Analysis |
Composite
| Benchmark | xhigh | max | not stated | Source | Trend |
|---|---|---|---|---|---|
| Epoch Capabilities Indexnot in index | – | 156.4#22 | – | Epoch AI Benchmarking Hub | |
| LiveBench | – | 78.5%#16 | – | LiveBench | |
| AA Intelligence Index v4.3.2 | – | 45.4#33 | – | Artificial Analysis | |
| Vals Indexnot in index | – | 48.3#28 | – | Vals AI | |
| Vals Multimodal Indexnot in index | – | 65.4%#12 | – | Vals AI |
Where it wins
Benchmarks where this configuration ranks in the top five of every configuration measured.
Where to run it
Every provider serving this model through OpenRouter, with throughput and first-token latency measured on live traffic over the last 30 minutes and each provider’s own price. Purple marks the best in each column.
| Provider | Speed | First token | Input $/M | Output $/M | Context | Quantisation |
|---|---|---|---|---|---|---|
| Alibaba Cloud Int. | 38 tok/s | 1.78 s | $2.00 | $6.00 | 1M | – |
| Alibaba Cloud Int. | 34 tok/s | 2.70 s | $2.00 | $6.00 | 1M | – |
What a task costs
Estimates from list price, output speed and time to first answer for the best configuration. “With caching” assumes three-quarters of the input is served from the prompt cache at $0.250 per 1M. Reasoning tokens are not modelled.
| Workload | Tokens in / out | Cost | With caching | Time |
|---|---|---|---|---|
| Chat reply | 400 / 300 | $0.0026 | $0.0021 | 1.1 min |
| Summarise a 30-page report | 12,000 / 600 | $0.028 | $0.012 | 1.3 min |
| Code edit | 6,000 / 1,500 | $0.021 | $0.013 | 1.7 min |
| Agentic coding session | 60,000 / 4,000 | $0.144 | $0.065 | 2.8 min |
| Structured extraction | 2,000 / 200 | $0.0052 | $0.0026 | 1.1 min |
See also
Data as of 11 Oct 2026. Compare these configurations.
Cite as: BenchLeader, “Qwen3.8 Max: benchmarks, pricing, speed and rank”, https://www.benchleader.com/models/qwen3-8-max, data as of 11 Oct 2026.