Model comparison
DeepSeek-V3.2-Exp vs Qwen3.6 27B
DeepSeek-V3.2-Exp is the stronger model overall, scoring 44.3 to 42.2 on the Noometry Index.
Last verified . 8 shared benchmarks.
DeepSeek
Alibaba (Qwen)Summary
- They share 8 benchmarks with published results for both. DeepSeek-V3.2-Exp scores higher in 2 categories and Qwen3.6 27B in 3 categories; 4 gaps are clear of the uncertainty.
- The widest gap is in writing & preference, where DeepSeek-V3.2-Exp leads 62.4 to 50.3.
- The biggest single-benchmark swing is DTBench: 87.7% for DeepSeek-V3.2-Exp and 78.1% for Qwen3.6 27B.
- DeepSeek-V3.2-Exp is cheaper at $0.26 / $0.38 per million input/output tokens, against $0.60 / $3.60 for Qwen3.6 27B.
- Qwen3.6 27B accepts more context: 262K tokens versus 164K.
Side by side
| DeepSeek-V3.2-Exp | Qwen3.6 27B | |
|---|---|---|
| Provider | DeepSeek | Alibaba (Qwen) |
| Noometry Index | 44.3 | 42.2 |
| Released | 2025-09-29 | 2026-04-22 |
| Weights | Open | Open |
| Context window | 164K | 262K |
| Max output | 66K | 66K |
| Input $ / M tokens | $0.26 | $0.60 |
| Output $ / M tokens | $0.38 | $3.60 |
| Results tracked | 49 | 11 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding DeepSeek-V3.2-Exp leads
DeepSeek-V3.2-Exp: 46.5 (#65), Qwen3.6 27B: 39.1 (#163)
| Benchmark | DeepSeek-V3.2-Exp | Qwen3.6 27B |
|---|---|---|
| SciCode | 38.9% | 37.3% |
| SWE-bench Verified (bash only) | 70% | — |
| Aider Polyglot | 74.2% | — |
| LMArena WebDev | 1362 | — |
| SWE-bench Multilingual | 59% | — |
| WeirdML | 39.5% | — |
| LMArena Coding | 1454 | — |
Agentic & Tool Use Not comparable
DeepSeek-V3.2-Exp: 32.7 (#59), Qwen3.6 27B: —
| Benchmark | DeepSeek-V3.2-Exp | Qwen3.6 27B |
|---|---|---|
| Terminal-Bench | 39.6% | — |
| APEX-Agents | 21.3% | — |
| Berkeley Function Calling Leaderboard | 56.7% | — |
| TheAgentCompany | 42.9% | — |
| Vending-Bench 2 | 1,034 | — |
Reasoning Qwen3.6 27B leads
DeepSeek-V3.2-Exp: 22.1 (#208), Qwen3.6 27B: 25.0 (#153)
| Benchmark | DeepSeek-V3.2-Exp | Qwen3.6 27B |
|---|---|---|
| CritPt | 2.9% | 0.9% |
| Chess Puzzles | 14% | 22% |
| DTBench | 87.7% | 78.1% |
| LMCA | 29.1% | 34.5% |
| Epoch Capabilities Index | 146.27 | 146.5 |
| ARC-AGI-2 | 4% | — |
| Kagi LLM Benchmark | 52.2% | — |
| NYT Connections (extended) | 36.7% | — |
| ARC-AGI-1 | 57% | — |
| Thematic Generalization | 65% | — |
| LMArena Hard Prompts | 1434 | — |
| Mystery Game Puzzles | — | 7% |
Math Qwen3.6 27B leads
DeepSeek-V3.2-Exp: 41.7 (#87), Qwen3.6 27B: 48.5 (#62)
| Benchmark | DeepSeek-V3.2-Exp | Qwen3.6 27B |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 87.8% | 91.1% |
| FrontierMath (Tiers 1-3) | — | 35.1% |
| MathArena Final-Answer Competitions | 57.7% | — |
| ProofBench | 8% | — |
| LMArena Math | 1435 | — |
| FrontierMath (Feb 2025 set) | 22.1% | — |
| FrontierMath Tier 4 (v1) | 2.1% | — |
Knowledge Too close to call
DeepSeek-V3.2-Exp: 51.7 (#66), Qwen3.6 27B: 52.4 (#63)
| Benchmark | DeepSeek-V3.2-Exp | Qwen3.6 27B |
|---|---|---|
| GPQA Diamond | 83.4% | 85.9% |
| Vectara Hallucination Rate | 5.3% | — |
| LMArena Expert | 1436 | — |
Multilingual Not comparable
DeepSeek-V3.2-Exp: 52.2 (#90), Qwen3.6 27B: —
| Benchmark | DeepSeek-V3.2-Exp | Qwen3.6 27B |
|---|---|---|
| LMArena Non-English | 1409 | — |
| LMArena Chinese | 1461 | — |
| LMArena French | 1433 | — |
| LMArena German | 1440 | — |
| LMArena Japanese | 1374 | — |
| LMArena Korean | 1371 | — |
| LMArena Russian | 1424 | — |
| LMArena Spanish | 1440 | — |
Instruction Following Not comparable
DeepSeek-V3.2-Exp: 74.5 (#93), Qwen3.6 27B: —
| Benchmark | DeepSeek-V3.2-Exp | Qwen3.6 27B |
|---|---|---|
| LMArena Instruction Following | 1413 | — |
Long Context Not comparable
DeepSeek-V3.2-Exp: 47.6 (#16), Qwen3.6 27B: —
| Benchmark | DeepSeek-V3.2-Exp | Qwen3.6 27B |
|---|---|---|
| Fiction.LiveBench | 83.3% | — |
| CL-bench | 13.2% | — |
| CL-bench Life | 9.5% | — |
| LMArena Longer Query | 1428 | — |
Writing & Preference DeepSeek-V3.2-Exp leads
DeepSeek-V3.2-Exp: 62.4 (#77), Qwen3.6 27B: 50.3 (#181)
| Benchmark | DeepSeek-V3.2-Exp | Qwen3.6 27B |
|---|---|---|
| LMArena Text | 1425 | — |
| LMArena Creative Writing | 1403 | — |
| EQ-Bench Creative Writing | 1515 | — |
| EQ-Bench 4 | — | 1026 |
| LMArena Multi-Turn | 1427 | — |
Frequently asked questions
Is DeepSeek-V3.2-Exp better than Qwen3.6 27B?
DeepSeek-V3.2-Exp is the stronger model overall, scoring 44.3 to 42.2 on the Noometry Index.
Which is cheaper, DeepSeek-V3.2-Exp or Qwen3.6 27B?
DeepSeek-V3.2-Exp is cheaper. It lists at $0.26 per million input tokens and $0.38 per million output tokens; Qwen3.6 27B lists at $0.60 and $3.60.
Is DeepSeek-V3.2-Exp or Qwen3.6 27B better for coding?
DeepSeek-V3.2-Exp scores higher on coding benchmarks: 46.5 versus 39.1 in the Noometry coding category.
Which has the bigger context window?
Qwen3.6 27B does, with 262K tokens against 164K.
How many benchmarks do DeepSeek-V3.2-Exp and Qwen3.6 27B share?
8 benchmarks have published results for both models. DeepSeek-V3.2-Exp has 49 scored results on Noometry and Qwen3.6 27B has 11.