Sign Up

Decision support

Qwen3.6 Max Preview vs Qwen3.6 27B

Compare canonical objects directly. The useful dimensions change with the object type, so model and hardware comparisons stay separate.

ModelsHardware
Qwen3.6 Max Preview

Alibaba

Sources:models.dev

Qwen3.6 27B

Qwen

Sources:models.devOpenRouter

Benchmark results

Head to head on 2 benchmarks both models report · 1 tie

  • Qwen3.6 Max Preview1lead
  • Qwen3.6 27B0leads

  • Arena Text

    Arena rating · Higher is better

    • Qwen3.6 Max Preview
      1460
      Arena #62CI 1451 to 1468
    • Qwen3.6 27B
      Not reported
    Runs and sources
    Qwen3.6 Max Preview
    • default146095% CI 1451 to 1468 · 5,386 votes
    Qwen3.6 27B

    Not reported

    Source:Arena Text leaderboard, 2026-11-09

  • Arena WebDev

    Arena rating · Higher is better

    • Qwen3.6 Max Preview
      1482
      Arena #75CI 1470 to 1495
    • Qwen3.6 27B
      Not reported
    Runs and sources
    Qwen3.6 Max Preview
    • default148295% CI 1470 to 1495 · 2,674 votes
    Qwen3.6 27B

    Not reported

    Source:Arena WebDev leaderboard, 2026-11-10

  • FrontierMath-2025-02-28-Private

    Best published score · Higher is better

    • Qwen3.6 Max Preview
      23.1%
      #20 of 58
    • Qwen3.6 27B
      Not reported
    Runs and sources
    Qwen3.6 Max Preview
    • default23.1%
    Qwen3.6 27B

    Not reported

    Source:Epoch AI evaluation

  • FrontierMath-Tier-4-2025-07-01-Private

    Best published score · Higher is better

    • Qwen3.6 Max Preview
      4.2%
      #18 of 46
    • Qwen3.6 27B
      Not reported
    Runs and sources
    Qwen3.6 Max Preview
    • default4.2%
    Qwen3.6 27B

    Not reported

    Source:Epoch AI evaluation

  • FrontierMath-Tiers-1-3-v2-Private

    Best published score · Higher is better

    • Qwen3.6 Max Preview
      Not reported
    • Qwen3.6 27B
      35.1%
      #63 of 87
    Runs and sources
    Qwen3.6 Max Preview

    Not reported

    Qwen3.6 27B
    • none34.0%
    • default35.1%
    Source:Epoch AI evaluation

  • GPQA diamond

    Best published score · Higher is better

    • Qwen3.6 Max Preview
      Best87.4%
      +1.5 pts ahead · #59 of 161
    • Qwen3.6 27B
      85.9%
      #67 of 161
    Runs and sources
    Qwen3.6 Max Preview
    • default87.4%
    Qwen3.6 27B
    • none84.8%
    • default85.9%
    Source:Epoch AI evaluation

  • OTIS Mock AIME 2024-2025

    Best published score · Higher is better

    Tie
    • Qwen3.6 Max Preview
      91.1%
      #54 of 158
    • Qwen3.6 27B
      91.1%
      #54 of 158
    Runs and sources
    Qwen3.6 Max Preview
    • default91.1%
    Qwen3.6 27B
    • none66.7%
    • default91.1%
    Source:Epoch AI evaluation

  • SimpleQA Verified

    Best published score · Higher is better

    • Qwen3.6 Max Preview
      52.0%
      #26 of 87
    • Qwen3.6 27B
      Not reported
    Runs and sources
    Qwen3.6 Max Preview
    • default52.0%
    Qwen3.6 27B

    Not reported

    Source:Epoch AI evaluation

  • SWE-bench Verified

    Best published score · Higher is better

    • Qwen3.6 Max Preview
      76.7%
      #7 of 35
    • Qwen3.6 27B
      Not reported
    Runs and sources
    Qwen3.6 Max Preview
    • default76.7%
    Qwen3.6 27B

    Not reported

    Source:Epoch AI evaluation

Each model's best published result is shown; "Not reported" means no matched evaluation, not a score of zero. Arena and DeepSWE ranks are the leaderboards' own; other ranks are among catalogue models with a score. Results from different harnesses or settings are not directly comparable.

Benchmark compilation: Epoch AI, Capabilities & benchmarking (CC BY attribution). Scores are converted to percentages and model names matched to the Vertical Tabs catalog. Individual evaluations are credited to their original sources. External results may carry separate licences and evaluation configurations; do not interpret scores from different harnesses as directly comparable.

Arena ratings: Arena (lmarena-ai/leaderboard-dataset), licensed CC BY 4.0. Ratings come from pairwise human votes and are only comparable within one leaderboard; the interval is Arena's 95% confidence interval. Model names are matched to the Vertical Tabs catalog.

Model specifications

Release date
April 20, 2026
April 22, 2026
Context window
262K
262K
Native context
—
256K
Max output tokens
65,536
81,920
Knowledge cutoff
2025-04
—
Open weights
Closed
Open weights
Parameter count
—
27B

API pricing · USD per 1M tokens

Input
—
$0.32
Output
—
$3.20
Cached input
—
—
Cache write
—
—

Capabilities and modalities

Input modalities
text
text, image, video
Output modalities
text
text
Tool calling
Supported
Supported
Reasoning option
Supported
Supported
Supported options
tools, reasoning, temperature
frequency_penalty, include_reasoning, logit_bias, logprobs, max_tokens, min_p, presence_penalty, reasoning, repetition_penalty, response_format, seed, stop, structured_outputs, temperature, tool_choice, tools, top_k, top_logprobs, top_p
Capabilities
—
—

Architecture and availability

Architecture
—
—
Model families
—
—
Tokenizer
—
Qwen3
Instruction format
—
—
License
—
—
Providers
0
6
Provider endpoints
0
6
Ollama variants
0
0
Metadata source
ai-research
ai-research

A dash means the field is not reported. Provider-specific performance and pricing may differ.

To give you the best experience, we use cookies for personalization, insights, and performance. Learn more.