Compare Models

Qwen 3.6 27B

Qwen 3.6 27B is an open-weight model from Alibaba, released April 22, 2026. Its best result is #40 of 91 on SAGE.

Release Date: Apr 22, 2026

Developer Alibaba  🇨🇳
Context Window 262k
Max Output Tokens 66k
Token Costs (in/out) $0.60/3.60
Weights Open
Input Modalities

Accuracy

52.00 %

Avg. Cost (In/Out)

$ 0.60 / $ 3.60

Latency

36 min 31 s

Vals Index
BenchmarksAccuracyRankings
MortgageTax

68.28%

±0.92
18/98
SAGE

45.58%

±3.43
40/91
TaxEval v2

71.26%

±0.88
80/145
Vibe Code Bench v1.1

11.94%

±3.41
90/110
SWE-bench

70.00%

±2.05
63/88
Terminal-Bench 2.0

44.94%

±5.30
26/67
Proprietary BenchmarksAcademic BenchmarksIndustry Partners
Vals
Default Provider : Alibaba
Temperature: 1
Top P: 0.95
Top K: 20
Max Output Tokens: 65,536

Updates

Apr 28, 2026

We evaluated Qwen 3.6 27B on our suite of benchmarks. Here are the key takeaways:

  • Alibaba’s Qwen 3.6 27B is on the Vals leaderboard: #22 of 46 on the Vals Index and #8 of 18 among open-weight models.
  • Head-to-head with the much larger Qwen 3.6 Plus, scores are within error margins of each other on the Vals Index, Vals Multimodal Index, as well as Case Law (v2), Mortgage Tax, SAGE, and Terminal-Bench 2.0.

All evals were run via the Alibaba API. We used temperature 1.0 by default, and 0.6 on coding benchmarks, which are the recommended settings on the Hugging Face model card. For agentic benchmarks we enabled preserve_thinking, following Qwen’s guidance.