AlibabaQwen3 VL 8B Instruct
Qwen3 VL 8B Instruct is an Alibaba open-weights reasoning model, released 14 Oct 2025. Its best configuration (thinking reasoning effort) ranks #586 of 785 on the BenchLeader Index at 42.7 ±1.4, in the lower half. The ± is the point: 277 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 (43) and lowest in agents & tools (37). At $0.375 per million tokens blended it is cheaper than most ranked models. Output speed of 39 tokens per second puts it in the slowest quarter, with a first token in 1.0 s. It has been measured at 2 reasoning-effort settings; this summary describes the best-scoring one, and the tabs above switch between them. Last measured 14 Oct 2025.stale: no new result in six months
- Blended price
- $0.375/M
- $0.250 in · $0.750 out
- Output speed
- 39 tok/s
- OpenRouter traffic, 7-day median; not yet measured by Artificial Analysis
- First answer
- 0.98 s
- Context
- 262k
- Full answer
- –
- Cost per run
- –
- Released
- 14 Oct 2025
- Overall index43
- Reasoning41
- Agents & tools37
- Knowledge38
- Instruction following43
- Multimodal39
- Long context41
- Composite39
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 |
|---|---|---|---|---|---|---|
| thinkingbest | 42.7 | #586 | 39 tok/s | 0.98 s | $0.0003 | Agents & tools 37 · Composite 39 · Instruction following 43 · Knowledge 38 · Long context 41 · Multimodal 39 · Reasoning 41 |
| not stated | 40.0 | #667 | 39 tok/s | 0.98 s | $0.0003 | Agents & tools 36 · Composite 38 · Instruction following 36 · Knowledge 38 · Long context 32 · Multimodal 30 · Reasoning 41 |
How its index has moved
9 Sept 2026 to 23 Oct 2026- thinking
- 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 | thinking | not stated | Source |
|---|---|---|---|
| GPQA Diamond (AA)not in index | 57.9%#388 | 42.7%#464 | Artificial Analysis |
| Humanity's Last Exam (AA)not in index | 3.9%#559 | 2.7%#611 | Artificial Analysis |
| CritPt | 0.3%#271 | 0.0%#315 | Artificial Analysis |
Agents & tools
| Benchmark | thinking | not stated | Source |
|---|---|---|---|
| Terminal-Bench Hard | 3.8%#298 | 2.3%#320 | Artificial Analysis |
| τ²-Bench Telecom (AA)not in index | 22.5%#311 | 29.2%#263 | Artificial Analysis |
Knowledge
| Benchmark | thinking | not stated | Source |
|---|---|---|---|
| AA-Omniscience | -52.3#407 | -51.9#405 | Artificial Analysis |
| AA-Omniscience: accuracynot in index | 20.4%#299 | 20.5%#297 | Artificial Analysis |
| AA-Omniscience: non-hallucinationnot in index | 8.6%#454 | 9.0%#442 | Artificial Analysis |
Instruction following
| Benchmark | thinking | not stated | Source |
|---|---|---|---|
| IFBench | 39.9%#262 | 32.3%#343 | Artificial Analysis |
Multimodal
| Benchmark | thinking | not stated | Source |
|---|---|---|---|
| MMMU-Pro | 56.6%#223 | 47.3%#250 | Artificial Analysis |
Long context
| Benchmark | thinking | not stated | Source |
|---|---|---|---|
| AA-LCR | 33.3%#381 | 16.7%#448 | Artificial Analysis |
Composite
| Benchmark | thinking | not stated | Source |
|---|---|---|---|
| AA Intelligence Index v4.3.2 | 8.2#444 | 7.3#486 | 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 | 39 tok/s | 0.98 s | $0.250 | $0.750 | 262k | bf16 |
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 | $0.0003 | – | 8.7 s |
| Summarize a 30-page report | 12,000 / 600 | $0.0035 | – | 16.4 s |
| Code edit | 6,000 / 1,500 | $0.0026 | – | 39.4 s |
| Agentic coding session | 60,000 / 4,000 | $0.018 | – | 1.7 min |
| Structured extraction | 2,000 / 200 | $0.0007 | – | 6.1 s |
More from Alibaba
See also
Questions people ask
- What are Qwen3 VL 8B Instruct's benchmark scores?
- As of 25 Oct 2026, Qwen3 VL 8B Instruct (thinking effort) scores 42.7 ±1.4 on the BenchLeader Index, strongest in instruction following (43) and weakest in agents & tools (37). That rests on 7 independent benchmark results, and every score on this page links to its publisher.
- How does Qwen3 VL 8B Instruct rank?
- Qwen3 VL 8B Instruct (thinking effort) is #586 of 785 ranked configurations on 25 Oct 2026. 277 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 VL 8B Instruct cost per 1M tokens?
- As of 25 Oct 2026, Qwen3 VL 8B Instruct costs $0.250 per 1M input tokens and $0.750 per 1M output tokens, $0.375 blended (three input tokens per output token). The list price last changed on 13 Oct 2026.
- When was Qwen3 VL 8B Instruct released?
- Alibaba released Qwen3 VL 8B Instruct on 14 Oct 2025. BenchLeader has results for it at 2 reasoning-effort settings.
Data as of 25 Oct 2026. Compare these configurations.
Cite as: BenchLeader, “Qwen3 VL 8B Instruct: benchmarks, pricing, speed and rank”, data as of 25 Oct 2026, https://www.benchleader.com/models/qwen3-vl-8b.