BenchLeader

Qwen3-4B-Instruct-2507

Reasoning effort
Quality#250 of 339
42.0Index
Below the median
−8.2 on the median

As of 16 Nov 2026, Qwen3-4B-Instruct-2507 by Alibaba scores 42.0 on the BenchLeader Index (608 ranked configurations), with its best results in maths (46) and weakest in long context (29). Released 6 Aug 2025.

Key facts (as of 16 Nov 2026)

  • Qwen3-4B-Instruct-2507 was released on 6 Aug 2025 (Alibaba).

In detail. Qwen3-4B-Instruct-2507 is an Alibaba open-weights model, released 6 Aug 2025. Its best configuration ranks #504 of 608 on the BenchLeader Index at 42.0 ±5.1, in the lower half. The ± is the point: 196 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 (46) and lowest in long context (29). 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 6 Aug 2025.stale: no new result in six months

Blended price
–
Output speed
–
First answer
–
Context
262k
via Artificial Analysis
Cost per task
–
Tokens per task
–
Released
6 Aug 2025
How it scores by categoryDashed line = average model (50). One step of 15 = one standard deviation.
  1. Overall index42
  2. Reasoning35
  3. Agents & tools44
  4. Maths46
  5. Instruction following37
  6. Long context29

Embed this rank

More badge styles
[![Qwen3-4B-Instruct-2507 on BenchLeader: #244 overall](https://www.benchleader.com/badge/qwen3-4b-2507.svg?style=compact)](https://www.benchleader.com/models/qwen3-4b-2507)

Paste it at the top of README.md, beside your other badges. The badge redraws with every data update and links back to this page.

Get an email when Alibaba ships a new model

And once when it gets ranked, plus the weekly review on Mondays.

Choose providers and emails

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.

EffortIndexRankSpeedFirst answerChat reply costCategories
thinking40.2#533–––Agents & tools 35 · Instruction following 51 · Long context 42 · Maths 13
not statedbest42.0#504–––Agents & tools 44 · Instruction following 37 · Long context 29 · Maths 46 · Reasoning 35

How its index has moved

10 Sep 2026 to 16 Nov 2026
354147
  • not stated
  • thinking

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.

Metadata

Benchmarkthinkingnot statedSource
AA Openness Indexnot in index44.4#9744.4#97Artificial Analysis

Reasoning

Benchmarkthinkingnot statedSource
GPQA Diamondeffort not stated–45.8%#249Epoch AI Benchmarking Hub
GPQA Diamond (AA)not in index66.7%#33551.7%#424Artificial Analysis
Humanity's Last Exam (AA)not in index6.2%#4274.5%#521Artificial Analysis
CritPtnot in index0.0%#3370.0%#337Artificial Analysis

Coding

Benchmarkthinkingnot statedSource
LiveCodeBench (AA)not in index64.1%#9537.7%#176Artificial Analysis

Agents & tools

Benchmarkthinkingnot statedSource
Terminal-Bench Hard1.5%#3434.5%#289Artificial Analysis
τ²-Bench Telecom (AA)not in index25.4%#29626.6%#285Artificial Analysis
BFCL Overalleffort not stated–35.7%#41Berkeley Function Calling Leaderboard

Maths

Benchmarkthinkingnot statedSource
OTIS Mock AIMEeffort not stated–52.2%#193Epoch AI Benchmarking Hub
AIME 202682.5%#30–MathArena
HMMT February 202653.0%#31–MathArena
MathArena Apex2.1%#29–MathArena
AIME 2025 (AA)not in index82.7%#5452.3%#132Artificial Analysis

Knowledge

Benchmarknot statedSource
Global-MMLU-Lite (AA)not in index67.7%#104Artificial Analysis

Instruction following

Benchmarkthinkingnot statedSource
IFBench49.8%#17733.5%#331Artificial Analysis

Long context

Benchmarkthinkingnot statedSource
AA-LCR v1.137.3%#39211.3%#487Artificial Analysis

Composite

Benchmarkthinkingnot statedSource
AA Intelligence Index v4.3.2not in index8.8#4346.7#518Artificial Analysis
AA Progress Index v1not in index206.4#398173.9#494Artificial Analysis

More from Alibaba

New this month

See also

Embed or cite

Show where Qwen3-4B-Instruct-2507 ranks on a model card or README. The badge redraws itself with every daily update, carries the data date and links back to this page.

[![Qwen3-4B-Instruct-2507 on BenchLeader: #244 overall](https://www.benchleader.com/badge/qwen3-4b-2507.svg)](https://www.benchleader.com/models/qwen3-4b-2507)

Cite this

The ranking changes daily, so a reference carries the data date. APA style:

BenchLeader. (2026, November 16). Qwen3-4B-Instruct-2507 on the BenchLeader Index (method v1.4) [Data set]. Retrieved November 16, 2026, from https://www.benchleader.com/models/qwen3-4b-2507

BibTeX:

@misc{benchleader_qwen3_4b_2507_20261116,
  author       = {{BenchLeader}},
  title        = {{Qwen3-4B-Instruct-2507 on the BenchLeader Index (method v1.4)}},
  year         = {2026},
  month        = nov,
  howpublished = {\url{https://www.benchleader.com/models/qwen3-4b-2507}},
  note         = {Data as of 2026-11-16}
}

Every badge style, the embeddable top-10 widget and how to cite the Index

Questions people ask

What are Qwen3-4B-Instruct-2507's benchmark scores?
As of 16 Nov 2026, Qwen3-4B-Instruct-2507 scores 42.0 ±5.1 on the BenchLeader Index, strongest in maths (46) and weakest in long context (29). That rests on independent benchmark results only, and every score on this page links to its publisher.
How does Qwen3-4B-Instruct-2507 rank?
Qwen3-4B-Instruct-2507 is #504 of 608 ranked configurations on 16 Nov 2026. 196 other configurations score within its margin, so the exact position matters less than the band. The leader is Claude Fable 5.2 (max effort) at 72.7.
When was Qwen3-4B-Instruct-2507 released?
Alibaba released Qwen3-4B-Instruct-2507 on 6 Aug 2025. BenchLeader has results for it at 2 reasoning-effort settings.

Data as of 16 Nov 2026. Compare these configurations.

Cite as: BenchLeader, “Qwen3-4B-Instruct-2507: benchmarks, pricing, speed and rank”, data as of 16 Nov 2026, https://www.benchleader.com/models/qwen3-4b-2507.