Skip to main content

Model comparison

Qwen3.5-27B vs ZAYA1-8B

Data verified

Head-to-head evidence from 3 shared benchmark results across 2 categories. Overall scores shown here use the public BenchAlign v5 ranking lane.

60.7/100
No comparison
Zyphra
N/A
2 category wins0 category wins

Public leaderboard positions: Qwen3.5-27B #45 (Supported); ZAYA1-8B unranked (Not scored). Intervals and evidence labels describe ranking uncertainty, not a guarantee for a specific workload.

Evidence parity. Qwen3.5-27B and ZAYA1-8B share 3 comparable benchmark results. 2 of 8 categories are comparable. 25 results are unique to Qwen3.5-27B; 8 to ZAYA1-8B.

Updated July 23, 2026
Shared results
3
Qwen3.5-27B only
25
ZAYA1-8B only
8
Comparable categories
2 / 8

Treat this as a split decision. Qwen3.5-27B makes more sense if instruction following is the priority or you need the larger 262K context window; ZAYA1-8B is the better fit if its strengths line up with your actual workload.

Confidence note. This is a partial-evidence comparison with 3 shared benchmark results across 2 evidence categories; 2 of 8 categories currently have scoreable aggregates for both models. Treat the verdict as directional until coverage is more balanced.

Why this result

Qwen3.5-27B and ZAYA1-8B finish on the same BenchAlign overall score, so this is less about a single winner and more about where the edge shows up. The BenchAlign headline says tie; the benchmark table is where the real choice happens.

Qwen3.5-27B gives you the larger context window at 262K, compared with 131K for ZAYA1-8B.

Category breakdown

Exact category averages are shown below. Not measured means BenchLM does not have enough sourced public coverage for that model and category.

Category scores and score margins for Qwen3.5-27B and ZAYA1-8B
CategoryQwen3.5-27BΔZAYA1-8B
Inst. FollowingQwen3.5-27B95.0Margin 30.9ZAYA1-8B64.1
KnowledgeQwen3.5-27B82.7Margin 9.1ZAYA1-8B73.6
AgenticQwen3.5-27B52.0MarginNo overlapZAYA1-8BNot measured
CodingQwen3.5-27B64.9MarginNo overlapZAYA1-8BNot measured
ReasoningQwen3.5-27B60.6MarginNo overlapZAYA1-8BNot measured
MathQwen3.5-27BNot measuredMarginNo overlapZAYA1-8B80.4
MultilingualQwen3.5-27B82.2MarginNo overlapZAYA1-8BNot measured

Decisive benchmark drivers

The largest measured benchmark gaps in this matchup, with exact reported values.

More
A · Qwen3.5-27BB · ZAYA1-8B
  1. GPQA

    Knowledge
    Source ↗
    A 85.5%B 71%
    Winner: Qwen3.5-27BΔ 14.5
    GPQA: Qwen3.5-27B scored 85.5%; ZAYA1-8B scored 71%. Qwen3.5-27B wins this benchmark.
  2. MMLU-Pro

    Knowledge
    Source ↗
    A 86.1%B 74.2%
    Winner: Qwen3.5-27BΔ 11.9
    MMLU-Pro: Qwen3.5-27B scored 86.1%; ZAYA1-8B scored 74.2%. Qwen3.5-27B wins this benchmark.
  3. IFEval

    Inst. Following
    Source ↗
    A 95%B 85.6%
    Winner: Qwen3.5-27BΔ 9.4
    IFEval: Qwen3.5-27B scored 95%; ZAYA1-8B scored 85.6%. Qwen3.5-27B wins this benchmark.

Operational comparison

Runtime and commercial metrics are compared only when both models have a complete sourced value.

MetricQwen3.5-27BZAYA1-8BComparison
Input / output priceUSD per 1M tokensQwen3.5-27B$0 input / $0 outputZAYA1-8B$0 input / $0 outputListed prices are equal.
Generation speedtokens per secondQwen3.5-27BNot availableZAYA1-8BNot availableA complete speed comparison is not available.
First-answer latencyseconds to first tokenQwen3.5-27BNot availableZAYA1-8BNot availableA complete latency comparison is not available.
Context windowmaximum listed tokensQwen3.5-27B262KZAYA1-8B131KQwen3.5-27B lists the larger context window.

Benchmark Deep Dive

Agentic
BenchmarkQwen3.5-27BZAYA1-8BResult
Terminal-Bench 2.0Source 41.6%Not comparable
BrowseCompSource 61%Not comparable
OSWorld-VerifiedSource 56.2%Not comparable
τ²-bench resultsSource 93.9%Not comparable
Gert LabsSource 39.41%Not comparable
BFCL v4Source 39.2%Not comparable
Coding
BenchmarkQwen3.5-27BZAYA1-8BResult
SWE-bench VerifiedSource 72.4%Not comparable
SWE-RebenchSource 58.9%Not comparable
AA-SciCodeSource 39.5%Not comparable
LiveCodeBench v6Source 65.8%Not comparable
Reasoning
BenchmarkQwen3.5-27BZAYA1-8BResult
LongBench v2Source 60.6%Not comparable
AA-LCRSource 67.3%Not comparable
CritPtSource 0.9%Not comparable
KnowledgeQwen3.5-27B wins
BenchmarkQwen3.5-27BZAYA1-8BResult
MMLU-ProSource 86.1%74.2%Qwen3.5-27B leads
SuperGPQASource 65.6%Not comparable
GPQASource 85.5%71%Qwen3.5-27B leads
Artificial Analysis Intelligence IndexSource 33.8%Not comparable
AA-GPQA DiamondSource 85.8%Not comparable
AA-HLESource 22.2%Not comparable
AA-Omniscience IndexSource -42.0%Not comparable
AA-Omniscience AccuracySource 21.0%Not comparable
AA-Omniscience Hallucination RateSource 79.7%Not comparable
GPQA-DSource 71.0%Not comparable
Math
BenchmarkQwen3.5-27BZAYA1-8BResult
AIME26Source 89.1%Not comparable
HMMT Feb 2026Source 71.6%Not comparable
IMOAnswerBenchSource 59.3%Not comparable
ApexSource 32.2%Not comparable
Multilingual
BenchmarkQwen3.5-27BZAYA1-8BResult
MMLU-ProXSource 82.2%Not comparable
Multimodal
BenchmarkQwen3.5-27BZAYA1-8BResult
MMMUSource 82.3%Not comparable
MMVUSource 73.3%Not comparable
MathVisionSource 86.0%Not comparable
V*Source 93.7%Not comparable
AA-MMMU-ProSource 75.0%Not comparable
Inst. FollowingQwen3.5-27B wins
BenchmarkQwen3.5-27BZAYA1-8BResult
IFEvalSource 95%85.6%Qwen3.5-27B leads
AA-IFBenchSource 75.6%Not comparable
IFBenchSource 52.6%Not comparable
Frequently Asked Questions (3)

Which is better, Qwen3.5-27B or ZAYA1-8B?

Qwen3.5-27B and ZAYA1-8B are tied on the BenchAlign overall score, so the right pick depends on which category matters most for your use case.

Which is better for knowledge tasks, Qwen3.5-27B or ZAYA1-8B?

Qwen3.5-27B has the edge for knowledge tasks in this comparison, averaging 82.7 versus 73.6. Inside this category, GPQA is the benchmark that creates the most daylight between them.

Which is better for instruction following, Qwen3.5-27B or ZAYA1-8B?

Qwen3.5-27B has the edge for instruction following in this comparison, averaging 95 versus 64.1. Inside this category, IFEval is the benchmark that creates the most daylight between them.

Related Comparisons

Last updated: July 23, 2026

Choose a model with this week’s evidence

Join 2,000+ readers for ranking moves, pricing changes, and the claims that still need proof.

One email each week. Unsubscribe anytime.