AlibabaAuto-detectedQwen3 8
Best configuration ranks #38 of 610 on the BenchLeader Index at 63.9 ±3.3 (max reasoning effort). Last measured 8 Sept 2026. Released 2 Aug 2026.
- Blended price
- $2.67/M
- $1.78 in · $5.33 out
- Output speed
- 39 tok/s
- First answer
- 53 s
- first token 1.78 s
- Context
- 1M
- Overall index64
- Reasoning68
- Coding72
- Agents & tools56
- Knowledge62
- Human preference68
- Multimodal67
- Long context66
- Composite71
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. “Default” means the publisher did not say which setting was used.
| Effort | Index | Rank | Speed | First answer | Chat reply cost | Categories |
|---|---|---|---|---|---|---|
| xhigh | 58.0 | #134 | 37 tok/s | 1.78 s | – | Knowledge 56 · Maths 65 · Reasoning 67 |
| maxbest | 63.9 | #38 | 39 tok/s | 53 s | $0.0023 | Agents & tools 56 · Coding 72 · Composite 71 · Human preference 68 · Knowledge 62 · Long context 66 · Multimodal 67 · Reasoning 68 |
| default | – | – | 37 tok/s | 1.78 s | – | Coding 62 · Maths 62 |
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 | default | Source | Trend |
|---|---|---|---|---|---|
| GPQA Diamond | 92.7%#20 | – | – | Epoch AI Benchmarking Hub | |
| LMArena Hard Prompts | – | 1502#22 | – | LMArena | |
| LiveBench Reasoningnot in index | – | 88.2%#16 | – | LiveBench | |
| GPQA Diamond (AA)not in index | – | 92.7%#31 | – | Artificial Analysis | |
| Humanity's Last Exam (AA)not in index | – | 43.0%#33 | – | Artificial Analysis | |
| GPQA Diamond (Vals)not in index | – | 93.7%#6 | – | Vals AI |
Coding
| Benchmark | xhigh | max | default | Source | Trend |
|---|---|---|---|---|---|
| SciCode | – | – | 52.9%#46 | SciCode | |
| LMArena Coding | – | 1520#30 | – | LMArena | |
| LMArena WebDev | – | 1685#4 | – | LMArena | |
| LiveBench Codingnot in index | – | 72.9%#41 | – | LiveBench | |
| SciCode (AA)not in index | – | 53.2%#49 | – | Artificial Analysis | |
| LiveCodeBench | – | 87.8%#10 | – | Vals AI | |
| IOI | – | 73.0%#3 | – | Vals AI | |
| SWE-bench (Vals)not in index | – | 85.6%#17 | – | Vals AI |
Agents & tools
| Benchmark | xhigh | max | default | Source | Trend |
|---|---|---|---|---|---|
| LMArena Agent | – | 3.9#14 | – | LMArena | |
| LiveBench Agentic Codingnot in index | – | 64.7%#4 | – | LiveBench | |
| Terminal-Bench 2.1 (Vals) | – | 67.4%#27 | – | Vals AI |
Maths
| Benchmark | xhigh | max | default | Source | Trend |
|---|---|---|---|---|---|
| FrontierMath Tiers 1–3 | 74.7%#13 | – | – | Epoch AI Benchmarking Hub | |
| FrontierMath Tier 4 | 46.3%#20 | – | – | Epoch AI Benchmarking Hub | |
| OTIS Mock AIME | 100.0%#1 | – | – | Epoch AI Benchmarking Hub | |
| ProofBench | – | – | 58.0%#11 | Vals AI | |
| LiveBench Mathematicsnot in index | – | 91.3%#18 | – | LiveBench |
Knowledge
| Benchmark | xhigh | max | default | Source | Trend |
|---|---|---|---|---|---|
| SimpleQA Verified | 47.3%#28 | – | – | Epoch AI Benchmarking Hub | |
| LiveBench Data Analysisnot in index | – | 78.4%#16 | – | LiveBench | |
| AA-Omniscience | – | 3.4#79 | – | Artificial Analysis | |
| MMLU-Pro | – | 88.6%#18 | – | Vals AI | |
| LegalBench | – | 83.6%#46 | – | Vals AI | |
| CorpFin | – | 65.8%#26 | – | Vals AI | |
| TaxEval | – | 75.5%#17 | – | Vals AI |
Instruction following
| Benchmark | xhigh | max | default | Source | Trend |
|---|---|---|---|---|---|
| LiveBench Languagenot in index | – | 79.7%#25 | – | LiveBench |
Human preference
| Benchmark | xhigh | max | default | Source | Trend |
|---|---|---|---|---|---|
| LMArena Text | – | 1480#22 | – | LMArena |
Multimodal
| Benchmark | xhigh | max | default | Source | Trend |
|---|---|---|---|---|---|
| LMArena Vision | – | 1313#6 | – | LMArena | |
| MMMU-Pro | – | 82.3%#21 | – | Artificial Analysis |
Long context
| Benchmark | xhigh | max | default | Source | Trend |
|---|---|---|---|---|---|
| AA-LCR | – | 78.3%#77 | – | Artificial Analysis |
Composite
| Benchmark | xhigh | max | default | Source | Trend |
|---|---|---|---|---|---|
| Epoch Capabilities Indexnot in index | – | 156.6#14 | – | Epoch AI Benchmarking Hub | |
| LiveBench | – | 78.5%#10 | – | LiveBench | |
| AA Intelligence Index | – | 40.3#33 | – | Artificial Analysis | |
| Vals Indexnot in index | – | 51.8#26 | – | 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. | 37 tok/s | 1.78 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 $0.222 per 1M. Reasoning tokens are not modelled.
| Workload | Tokens in / out | Cost | With caching | Time |
|---|---|---|---|---|
| Chat reply | 400 / 300 | $0.0023 | $0.0018 | 1.0 min |
| Summarise a 30-page report | 12,000 / 600 | $0.025 | $0.011 | 1.1 min |
| Code edit | 6,000 / 1,500 | $0.019 | $0.012 | 1.5 min |
| Agentic coding session | 60,000 / 4,000 | $0.128 | $0.058 | 2.6 min |
| Structured extraction | 2,000 / 200 | $0.0046 | $0.0023 | 58.3 s |
See also
Data as of 9 Sept 2026. Compare these configurations.