AlibabaQwen3.5-9B
Qwen3.5-9B is an Alibaba open-weights reasoning model, released 24 Feb 2026. Its best configuration (thinking reasoning effort) ranks #410 of 785 on the BenchLeader Index at 49.0 ±7.3, in the lower half. The ± is the point: 424 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 instruction following (66) and lowest in knowledge (37). At $0.155 per million tokens blended it is among the cheapest fifth of ranked models. Output speed of 59 tokens per second puts it slower than most, with a first answer in 36.0 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 2 Mar 2026.stale: no new result in six months
- Blended price
- $0.155/M
- $0.140 in · $0.200 out
- Output speed
- 59 tok/s
- measured by Artificial Analysis
- First answer
- 36 s
- first token 0.77 s
- Context
- 262k
- Full answer
- 45 s
- median, reasoning included
- Cost per run
- $0.213
- one full Intelligence Index run
- Released
- 24 Feb 2026
- Overall index49
- Reasoning41
- Agents & tools41
- Knowledge37
- Instruction following66
- Multimodal52
- Long context60
- Composite42
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 | 46.4 | #488 | 79 tok/s | 0.75 s | $0.0001 | Agents & tools 50 · Composite 44 · Instruction following 41 · Knowledge 30 · Long context 47 · Maths 50 · Multimodal 50 · Reasoning 50 |
| thinkingbest | 49.0 | #410 | 59 tok/s | 36 s | $0.0001 | Agents & tools 41 · Composite 42 · Instruction following 66 · Knowledge 37 · Long context 60 · Multimodal 52 · Reasoning 41 |
| not stated | 40.1 | #662 | 70 tok/s | 0.77 s | $0.0001 | Agents & tools 24 · Coding 26 · Maths 37 · Reasoning 57 |
How its index has moved
9 Sept 2026 to 21 Oct 2026- thinking
- no reasoning
- not stated
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 | 78.9%#131 | – | 79.0%#130 | Epoch AI Benchmarking Hub | |
| GPQA Diamond (AA)not in index | 78.6%#208 | 80.6%#190 | – | Artificial Analysis | |
| Humanity's Last Exam (AA)not in index | 9.4%#336 | 14.9%#254 | – | Artificial Analysis | |
| CritPt | 0.6%#254 | 0.3%#271 | – | Artificial Analysis | |
| LMCAnot in indexeffort not stated | – | – | 24.5%#150 | Epoch AI Benchmarking Hub | |
| DTBenchnot in indexeffort not stated | – | – | 71.2%#155 | Epoch AI Benchmarking Hub |
Coding
| Benchmark | no reasoning | thinking | not stated | Source | Trend |
|---|---|---|---|---|---|
| SciCodeeffort not stated | – | – | 29.5%#198 | SciCode | |
| SciCode (AA)not in index | – | 29.5%#212 | – | Artificial Analysis | |
| Terminal-Bench 4.0 (AA)not in index | – | 0.5%#139 | – | Artificial Analysis | |
| Terminal-Bench 2.1 (AA)not in index | 21.4%#170 | 29.2%#157 | – | Artificial Analysis |
Agents & tools
| Benchmark | no reasoning | thinking | not stated | Source | Trend |
|---|---|---|---|---|---|
| Terminal-Bencheffort not stated | – | – | 9.2%#83 | Terminal-Bench | |
| Terminal-Bench Hard | 18.2%#171 | 24.2%#144 | – | Artificial Analysis | |
| τ²-Bench Telecom (AA)not in index | 85.1%#94 | 86.8%#83 | – | Artificial Analysis | |
| GDPval-AA v2.1not in index | – | 0.0%#124 | – | Artificial Analysis | |
| τ³-Banking (AA) | – | 7.0%#163 | – | Artificial Analysis | |
| AutomationBenchnot in index | – | 1.1%#194 | – | Artificial Analysis | |
| GDP.pdf (AA)not in index | – | 6.6%#150 | – | Artificial Analysis |
Maths
| Benchmark | no reasoning | not stated | Source | Trend |
|---|---|---|---|---|
| OTIS Mock AIME | 61.7%#162 | 42.8%#196 | Epoch AI Benchmarking Hub | |
| AIME 2026effort not statedpossible contamination | – | 92.5%#23 | MathArena | |
| HMMT February 2026effort not statedpossible contamination | – | 71.2%#29 | MathArena | |
| MathArena Apexeffort not statedpossible contamination | – | 0.5%#41 | MathArena |
Knowledge
| Benchmark | no reasoning | thinking | Source | Trend |
|---|---|---|---|---|
| AA-Omniscience | -70.7#517 | -53.5#420 | Artificial Analysis | |
| AA-Omniscience: accuracynot in index | 14.0%#446 | 16.4%#389 | Artificial Analysis | |
| AA-Omniscience: non-hallucinationnot in index | 1.6%#551 | 16.4%#327 | Artificial Analysis |
Instruction following
| Benchmark | no reasoning | thinking | Source | Trend |
|---|---|---|---|---|
| IFBench | 37.8%#289 | 66.7%#91 | Artificial Analysis |
Multimodal
| Benchmark | no reasoning | thinking | Source | Trend |
|---|---|---|---|---|
| MMMU-Pro | 66.8%#168 | 69.3%#153 | Artificial Analysis |
Long context
| Benchmark | no reasoning | thinking | Source | Trend |
|---|---|---|---|---|
| AA-LCR | 46.0%#333 | 70.0%#215 | Artificial Analysis |
Composite
| Benchmark | no reasoning | thinking | not stated | Source | Trend |
|---|---|---|---|---|---|
| Epoch Capabilities Indexnot in indexeffort not stated | – | – | 139.3#115 | Epoch AI Benchmarking Hub | |
| AA Intelligence Index v4.3.2 | 13.3#312 | 11.2#359 | – | 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 |
|---|---|---|---|---|---|---|
| Parasail | 101 tok/s | 0.45 s | $0.100 | $0.250 | 262k | bf16 |
| Together | 69 tok/s | 1.22 s | $0.170 | $0.250 | 262k | – |
| DeepInfra | 32 tok/s | 0.79 s | $0.100 | $0.150 | 262k | bf16 |
| SiliconFlow | 24 tok/s | 1.01 s | $0.100 | $0.150 | 262k | fp8 |
| Darkbloom | 22 tok/s | 2.08 s | $0.080 | $0.130 | 262k | fp4 |
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 at $0.060 per 1M. Reasoning tokens are not modelled.
| Workload | Tokens in / out | Cost | With caching | Time |
|---|---|---|---|---|
| Chat reply | 400 / 300 | $0.0001 | $0.0001 | 41.1 s |
| Summarize a 30-page report | 12,000 / 600 | $0.0018 | $0.0011 | 46.2 s |
| Code edit | 6,000 / 1,500 | $0.0011 | $0.0008 | 1.0 min |
| Agentic coding session | 60,000 / 4,000 | $0.0092 | $0.0056 | 1.7 min |
| Structured extraction | 2,000 / 200 | $0.0003 | $0.0002 | 39.4 s |
See also
Questions people ask
- What are Qwen3.5-9B's benchmark scores?
- As of 24 Oct 2026, Qwen3.5-9B (thinking effort) scores 49.0 ±7.3 on the BenchLeader Index, strongest in instruction following (66) and weakest in knowledge (37). That rests on 8 independent benchmark results, and every score on this page links to its publisher.
- How does Qwen3.5-9B rank?
- Qwen3.5-9B (thinking effort) is #410 of 785 ranked configurations on 24 Oct 2026. 424 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 Qwen3.5-9B cost per 1M tokens?
- As of 24 Oct 2026, Qwen3.5-9B costs $0.140 per 1M input tokens and $0.200 per 1M output tokens, $0.155 blended (three input tokens per output token). The list price last changed on 6 Oct 2026.
- When was Qwen3.5-9B released?
- Alibaba released Qwen3.5-9B on 24 Feb 2026. BenchLeader has results for it at 3 reasoning-effort settings.
Data as of 24 Oct 2026. Compare these configurations.
Cite as: BenchLeader, “Qwen3.5-9B: benchmarks, pricing, speed and rank”, data as of 24 Oct 2026, https://www.benchleader.com/models/qwen3-5-9b.