GPQA Diamond ↗85.86%
Qwen3.6 27B · GPQA diamond · Best score (across scorers) · % ↗
1 more reported result
84.85% · Qwen3.6 27B · GPQA diamond · Best score (across scorers) · % · Epoch AI · GPQA diamond ↗
Qwen / MODEL EVIDENCE
Qwen3.6 27B
15 benchmarks · 19 observations · 16 sources · released 2026-04-22
Scores at a glance
15 benchmarks · large dot: Qwen3.6 27B; small dots: closest models where reported
Qwen3.6 27B · FrontierMath-Tiers-1-3-v2-Private · Best score (across scorers) · % ↗
1 more reported result
34.04% · Qwen3.6 27B · FrontierMath-Tiers-1-3-v2-Private · Best score (across scorers) · % · Epoch AI · FrontierMath-Tiers-1-3-v2-Private ↗
Qwen3.6 27B · OTIS Mock AIME 2024-2025 · Best score (across scorers) · % ↗
1 more reported result
66.67% · Qwen3.6 27B · OTIS Mock AIME 2024-2025 · Best score (across scorers) · % · Epoch AI · OTIS Mock AIME 2024-2025 ↗
Qwen3.6 27B · Mystery Game Puzzles · Best score (across scorers) · % ↗
CritPt ↗0.86%
Qwen3.6 27B · CritPt · Accuracy · % ↗
LMCA ↗34.53%
Qwen3.6 27B · LMCA · Score · % ↗
DTBench ↗78.13%
Qwen3.6 27B · DTBench · Accuracy · % ↗
SciCode ↗37.27%
Qwen3.6 27B · SciCode · Score · % ↗
Epoch Capabilities Index (ECI) ↗146.23 ECI
Qwen3.6 27B · General ECI · 26 October 2026 fit ↗
CaseLaw v2 ↗53.16 %
Qwen 3.6 27B · CaseLaw · Accuracy ↗
CorpFin v2 ↗62.31 %
Qwen 3.6 27B · CorpFin · Accuracy ↗
MortgageTax & TaxEval v2 ↗68.28 %
Qwen 3.6 27B · MortgageTax · Accuracy ↗
1 more reported result
71.26 % · Qwen 3.6 27B · TaxEval · Accuracy · Vals AI · TaxEval ↗
Qwen 3.6 27B · SAGE · Accuracy ↗
Vibe Code Bench ↗11.94 %
Qwen 3.6 27B (OpenHands) · Vibe Code Bench · original task set ↗
Qwen 3.6 27B (Mini-SWE-agent) · SWE-bench Verified · Score ↗
Closest models by score
Similar positions on shared benchmark leaderboards. Smaller gaps are closer.
Claude Sonnet 4.6 →Anthropic
5 of 7 shared benchmark comparisons
Claude Opus 4.5 →Anthropic
5 of 7 shared benchmark comparisons
GPT-5.4 →OpenAI
5 of 7 shared benchmark comparisons
How similarity is measured
Mean absolute gap in percentile positions on shared benchmark settings; lower is closer. At least 60% overlap is required, with a minimum of three when available. Independent evaluations are used when there are at least three; otherwise published report comparisons are used. Repeated model observations within one setting are averaged before ranking. Writing-style traits and composite indices are excluded. This describes similarity, not an overall quality ranking.
Sources
Every plotted observation links to its evaluation. Reported configurations, benchmark versions and task subsets can differ; similarity describes shared scores, not training lift.
- SciCode Benchmark | Artificial Analysis ↗
- DTBench — Conceptual Reasoning Index ↗
- LMCA — Conceptual Reasoning Index ↗
- CritPt v1 ↗
- FrontierMath Tiers 1-3 (v2) ↗
- GPQA Diamond ↗
- Mystery Game Puzzles ↗
- OTIS Mock AIME 2024-2025 ↗
- AI benchmarks & capabilities ↗
- CaseLaw — evaluation methodology and results ↗
- CorpFin — evaluation methodology and results ↗
- MortgageTax — evaluation methodology and results ↗
- SAGE — evaluation methodology and results ↗
- SWE-bench Verified — evaluation methodology and results ↗
- TaxEval — evaluation methodology and results ↗
- Vibe Code Bench — evaluation methodology and results ↗
Names used in source tables
Qwen 3.6 27B · Qwen 3.6 27B (Mini-SWE-agent) · Qwen 3.6 27B (OpenHands) · Qwen3.6 27B