AlibabaQwen3 30B A3B
Qwen3 30B A3B is an Alibaba open-weights reasoning model, released 29 Apr 2025. Its best configuration ranks #432 of 785 on the BenchLeader Index at 48.3 ±3.7, in the lower half. The ± is the point: 296 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 maths (52) and lowest in coding (43). At $0.215 per million tokens blended it is cheaper than most ranked models. Output speed of 118 tokens per second puts it faster than most, with a first token in 0.5 s. It has been superseded by Qwen3 30B A3B 2507, which is the version to compare against today. 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 22 Oct 2026.
- Blended price
- $0.215/M
- $0.120 in · $0.500 out
- Output speed
- 118 tok/s
- OpenRouter traffic, 7-day median; not yet measured by Artificial Analysis
- First answer
- 0.50 s
- Context
- 41k
- Full answer
- –
- Cost per run
- –
- Released
- 29 Apr 2025
- Overall index48
- Reasoning47
- Coding43
- Maths52
- Human preference49
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 30B A3B 2507 | 30 Jul 2025 | 45.6 | #509 |
| Qwen3 30B A3Bthis page | 29 Apr 2025 | 48.3 | #432 |
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 |
|---|---|---|---|---|---|---|
| no reasoning | 38.7 | #705 | 145 tok/s | 0.59 s | $0.0002 | Agents & tools 40 · Composite 37 · Instruction following 36 · Knowledge 32 · Long context 23 · Maths 35 · Reasoning 40 |
| thinking | 41.1 | #641 | 130 tok/s | 16 s | $0.0002 | Agents & tools 36 · Composite 38 · Instruction following 44 · Knowledge 38 · Long context 23 · Reasoning 41 |
| not statedbest | 48.3 | #432 | 118 tok/s | 0.50 s | $0.0002 | Coding 43 · Human preference 49 · Maths 52 · Reasoning 47 |
How its index has moved
8 Sept 2026 to 23 Oct 2026- not stated
- thinking
- no reasoning
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
See all LLM benchmarks. 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 | no reasoning | thinking | not stated | Source | Trend |
|---|---|---|---|---|---|
| GPQA Diamond | 50.4%#214 | – | 61.7%#190 | Epoch AI Benchmarking Hub | |
| Arena Hard Promptseffort not stated | – | – | 1344#220 | Arena (LMArena) | |
| GPQA Diamond (AA)not in index | 51.5%#421 | 61.6%#364 | – | Artificial Analysis | |
| Humanity's Last Exam (AA)not in index | 4.6%#488 | 6.2%#405 | – | Artificial Analysis | |
| CritPt | 0.0%#315 | 0.0%#315 | – | Artificial Analysis | |
| LMCAnot in indexeffort not stated | – | – | 15.8%#171 | Epoch AI Benchmarking Hub | |
| DTBenchnot in indexeffort not stated | – | – | 60.3%#185 | Epoch AI Benchmarking Hub |
Coding
| Benchmark | not stated | Source | Trend |
|---|---|---|---|
| WeirdML | 29.8%#133 | WeirdML | |
| Arena Coding | 1385#211 | Arena (LMArena) |
Agents & tools
| Benchmark | no reasoning | thinking | Source | Trend |
|---|---|---|---|---|
| Terminal-Bench Hard | 6.8%#249 | 2.3%#320 | Artificial Analysis | |
| τ²-Bench Telecom (AA)not in index | 22.2%#314 | 26.0%#287 | Artificial Analysis |
Maths
| Benchmark | no reasoning | not stated | Source | Trend |
|---|---|---|---|---|
| OTIS Mock AIME | 25.6%#215 | 62.8%#158 | Epoch AI Benchmarking Hub | |
| Arena Mathseffort not stated | – | 1351#199 | Arena (LMArena) |
Knowledge
| Benchmark | no reasoning | thinking | Source | Trend |
|---|---|---|---|---|
| AA-Omniscience | -66.0#502 | -51.5#402 | Artificial Analysis | |
| AA-Omniscience: accuracynot in index | 12.0%#475 | 16.2%#397 | Artificial Analysis | |
| AA-Omniscience: non-hallucinationnot in index | 11.4%#389 | 19.3%#288 | Artificial Analysis |
Instruction following
| Benchmark | no reasoning | thinking | not stated | Source | Trend |
|---|---|---|---|---|---|
| IFBench | 31.9%#346 | 41.5%#246 | – | Artificial Analysis | |
| Arena Instruction Followingnot in indexeffort not stated | – | – | 1309#229 | Arena (LMArena) |
Human preference
| Benchmark | not stated | Source | Trend |
|---|---|---|---|
| Arena Text | 1327#226 | Arena (LMArena) | |
| Arena Creative Writingnot in index | 1283#235 | Arena (LMArena) | |
| Arena Multi-turnnot in index | 1319#227 | Arena (LMArena) | |
| Arena Longer Queriesnot in index | 1338#217 | Arena (LMArena) | |
| Lech Mazur Creative Writingnot in index | 75.3%#25 | Lech Mazur |
Long context
| Benchmark | no reasoning | thinking | Source | Trend |
|---|---|---|---|---|
| AA-LCR | 0.0%#493 | 0.0%#493 | Artificial Analysis |
Composite
| Benchmark | no reasoning | thinking | not stated | Source | Trend |
|---|---|---|---|---|---|
| Epoch Capabilities Indexnot in indexeffort not stated | – | – | 136.1#128 | Epoch AI Benchmarking Hub | |
| AA Intelligence Index v4.3.2 | 6.6#518 | 7.6#463 | – | Artificial Analysis |
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 |
|---|---|---|---|---|---|---|
| DeepInfra | 141 tok/s | 0.41 s | $0.120 | $0.500 | 41k | fp8 |
Price history
The price above, per 1M tokens, every day it changed.
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. Reasoning tokens are not modelled.
| Workload | Tokens in / out | Cost | With caching | Time |
|---|---|---|---|---|
| Chat reply | 400 / 300 | $0.0002 | – | 3.0 s |
| Summarize a 30-page report | 12,000 / 600 | $0.0017 | – | 5.6 s |
| Code edit | 6,000 / 1,500 | $0.0015 | – | 13.2 s |
| Agentic coding session | 60,000 / 4,000 | $0.0092 | – | 34.4 s |
| Structured extraction | 2,000 / 200 | $0.0003 | – | 2.2 s |
More from Alibaba
See also
Questions people ask
- What are Qwen3 30B A3B's benchmark scores?
- As of 24 Oct 2026, Qwen3 30B A3B scores 48.3 ±3.7 on the BenchLeader Index, strongest in maths (52) and weakest in coding (43). That rests on 7 independent benchmark results, and every score on this page links to its publisher.
- How does Qwen3 30B A3B rank?
- Qwen3 30B A3B is #432 of 785 ranked configurations on 24 Oct 2026. 296 other configurations score within its margin, so the exact position matters less than the band. The leader is Claude Opus 5.5 (max effort) at 72.6.
- How much does Qwen3 30B A3B cost per 1M tokens?
- As of 24 Oct 2026, Qwen3 30B A3B costs $0.120 per 1M input tokens and $0.500 per 1M output tokens, $0.215 blended (three input tokens per output token). The list price last changed on 19 Sept 2026.
- When was Qwen3 30B A3B released?
- Alibaba released Qwen3 30B A3B on 29 Apr 2025. BenchLeader has results for it at 3 reasoning-effort settings.
Data as of 24 Oct 2026. Compare these configurations.
Cite as: BenchLeader, “Qwen3 30B A3B: benchmarks, pricing, speed and rank”, data as of 24 Oct 2026, https://www.benchleader.com/models/qwen3-30b-a3b.