Models/Qwen3.5 2B

MDL

Qwen3.5 2B

Alibaba

Verified against Artificial Analysis · Jul 27, 2026

Capability

percentile index

Price /M

$0 / $0

in / out

Context

—

Avg score

25.9

8 benchmarks

AA Intelligence

6.9

index

Weights

Open

Released

Mar 2, 2026

About

Qwen3.5 2B is a open-weights model in the Qwen3.5 family from Alibaba. Benchmarked on 8 evals, averaging 25.9.

Benchmark scores · 8

avg 25.9 — higher is better

τ²-benchartificial-analysis · 2026-07

MATHartificial-analysis · 2026-07

GPQA Diamondartificial-analysis · 2026-07

32.5

IFBenchartificial-analysis · 2026-07

30.4

AA Long-Context Reasoningartificial-analysis · 2026-07

23.7

Terminal-Bench Hardartificial-analysis · 2026-07

3.8

SciCodeartificial-analysis · 2026-07

2.8

Humanity's Last Examartificial-analysis · 2026-07

2.1

The research behind this model

all papers mentioning it →

HRBench: Benchmarking and Understanding Thinking-Mode Switch Strategies in Hybrid-Reasoning LLMs▲ 15 on HF · May 27, 2026

Mentions matched by name in title/abstract — from the arXiv + HF daily corpus.

Details

PublisherAlibaba

FamilyQwen3.5

Parameters—

License—

WeightsOpen

ReleasedMar 2, 2026

Performance & price

Price in / out$0 / $0 per 1M

Coding index2.9

Median across API providers, via Artificial Analysis. Blended = 3:1 input:output. For reasoning models, latency includes thinking time. Methodology →

History

full ledger →

Capability over time

GPQA Diamond45.6 → 32.5

yesterday

MATHfirst seen at 43

yesterday

IFBench31.5 → 30.4

yesterday

Append-only ledger — every observed change to this model's numbers.

API providers

Artificial Analysis$0/M