Compare · ModelsLive · 2 picked · head to head
DeepSeek V3.2 vs Qwen3 Next 80B A3B Instruct
Side by side · benchmarks, pricing, and signals you can act on.
Winner summary
DeepSeek V3.2 wins on 14/18 benchmarks
DeepSeek V3.2 wins 14 of 18 shared benchmarks. Leads in speed · arena · coding.
Category leads
speed·DeepSeek V3.2arena·DeepSeek V3.2coding·DeepSeek V3.2reasoning·Qwen3 Next 80B A3B Instructlanguage·DeepSeek V3.2math·Qwen3 Next 80B A3B Instructknowledge·DeepSeek V3.2
Hype vs Reality
Attention vs performance
DeepSeek V3.2
#108 by perf·no signal
Qwen3 Next 80B A3B Instruct
#89 by perf·no signal
Best value
DeepSeek V3.2
2.1x better value than Qwen3 Next 80B A3B Instruct
DeepSeek V3.2
191.0 pts/$
$0.27/M
Qwen3 Next 80B A3B Instruct
91.4 pts/$
$0.60/M
Vendor risk
Mixed exposure
One or more vendors flagged
DeepSeek
$3.4B·Tier 1
Alibaba (Qwen)
$293.0B·Tier 1
Head to head
18 benchmarks · 2 models
DeepSeek V3.2Qwen3 Next 80B A3B Instruct
Artificial Analysis · Agentic Index
DeepSeek V3.2 leads by +38.7
Artificial Analysis Agentic Index · a composite score measuring how well a model performs in agentic workflows · multi-step tool use, planning, error recovery, and autonomous task completion. Aggregates results from multiple agentic benchmarks including SWE-bench, tool-use tests, and planning evaluations. The canonical single-number metric for "how good is this model as an agent?"
DeepSeek V3.2
52.9
Qwen3 Next 80B A3B Instruct
14.2
Artificial Analysis · Coding Index
DeepSeek V3.2 leads by +21.4
Artificial Analysis Coding Index · a composite score that aggregates performance across multiple coding benchmarks into a single index. Tracks code generation quality, debugging ability, multi-language competence, and real-world software engineering tasks. Used by Artificial Analysis to rank model coding capability in a normalized, comparable format. Useful for developers choosing between models for coding-heavy workloads.
DeepSeek V3.2
36.7
Qwen3 Next 80B A3B Instruct
15.3
Artificial Analysis · Quality Index
DeepSeek V3.2 leads by +28.0
DeepSeek V3.2
41.7
Qwen3 Next 80B A3B Instruct
13.7
Chatbot Arena Elo · Overall
DeepSeek V3.2 leads by +23.4
DeepSeek V3.2
1425.0
Qwen3 Next 80B A3B Instruct
1401.6
LiveBench · Agentic Coding
DeepSeek V3.2 leads by +36.7
DeepSeek V3.2
46.7
Qwen3 Next 80B A3B Instruct
10.0
LiveBench · Coding
DeepSeek V3.2 leads by +7.5
DeepSeek V3.2
75.7
Qwen3 Next 80B A3B Instruct
68.2
LiveBench · Data Analysis
Qwen3 Next 80B A3B Instruct leads by +4.8
DeepSeek V3.2
45.0
Qwen3 Next 80B A3B Instruct
49.8
LiveBench · If
DeepSeek V3.2 leads by +3.9
DeepSeek V3.2
23.1
Qwen3 Next 80B A3B Instruct
19.2
LiveBench · Language
Qwen3 Next 80B A3B Instruct leads by +2.1
DeepSeek V3.2
64.2
Qwen3 Next 80B A3B Instruct
66.3
LiveBench · Mathematics
Qwen3 Next 80B A3B Instruct leads by +6.2
DeepSeek V3.2
64.0
Qwen3 Next 80B A3B Instruct
70.2
LiveBench · Overall
DeepSeek V3.2 leads by +3.5
DeepSeek V3.2
51.8
Qwen3 Next 80B A3B Instruct
48.4
LiveBench · Reasoning
Qwen3 Next 80B A3B Instruct leads by +10.5
DeepSeek V3.2
44.3
Qwen3 Next 80B A3B Instruct
54.8
OpenCompass · AIME2025
DeepSeek V3.2 leads by +23.8
DeepSeek V3.2
93.0
Qwen3 Next 80B A3B Instruct
69.2
OpenCompass · GPQA-Diamond
DeepSeek V3.2 leads by +10.5
DeepSeek V3.2
84.6
Qwen3 Next 80B A3B Instruct
74.1
OpenCompass · HLE
DeepSeek V3.2 leads by +15.2
DeepSeek V3.2
23.2
Qwen3 Next 80B A3B Instruct
8.0
OpenCompass · IFEval
DeepSeek V3.2 leads by +2.1
DeepSeek V3.2
89.7
Qwen3 Next 80B A3B Instruct
87.6
OpenCompass · LiveCodeBenchV6
DeepSeek V3.2 leads by +20.6
DeepSeek V3.2
75.4
Qwen3 Next 80B A3B Instruct
54.8
OpenCompass · MMLU-Pro
DeepSeek V3.2 leads by +4.5
DeepSeek V3.2
85.8
Qwen3 Next 80B A3B Instruct
81.3
Full benchmark table
| Benchmark | DeepSeek V3.2 | Qwen3 Next 80B A3B Instruct |
|---|---|---|
Artificial Analysis · Agentic Index Artificial Analysis Agentic Index · a composite score measuring how well a model performs in agentic workflows · multi-step tool use, planning, error recovery, and autonomous task completion. Aggregates results from multiple agentic benchmarks including SWE-bench, tool-use tests, and planning evaluations. The canonical single-number metric for "how good is this model as an agent?" | 52.9 | 14.2 |
Artificial Analysis · Coding Index Artificial Analysis Coding Index · a composite score that aggregates performance across multiple coding benchmarks into a single index. Tracks code generation quality, debugging ability, multi-language competence, and real-world software engineering tasks. Used by Artificial Analysis to rank model coding capability in a normalized, comparable format. Useful for developers choosing between models for coding-heavy workloads. | 36.7 | 15.3 |
Artificial Analysis · Quality Index | 41.7 | 13.7 |
Chatbot Arena Elo · Overall | 1425.0 | 1401.6 |
LiveBench · Agentic Coding | 46.7 | 10.0 |
LiveBench · Coding | 75.7 | 68.2 |
LiveBench · Data Analysis | 45.0 | 49.8 |
LiveBench · If | 23.1 | 19.2 |
LiveBench · Language | 64.2 | 66.3 |
LiveBench · Mathematics | 64.0 | 70.2 |
LiveBench · Overall | 51.8 | 48.4 |
LiveBench · Reasoning | 44.3 | 54.8 |
OpenCompass · AIME2025 | 93.0 | 69.2 |
OpenCompass · GPQA-Diamond | 84.6 | 74.1 |
OpenCompass · HLE | 23.2 | 8.0 |
OpenCompass · IFEval | 89.7 | 87.6 |
OpenCompass · LiveCodeBenchV6 | 75.4 | 54.8 |
OpenCompass · MMLU-Pro | 85.8 | 81.3 |
Pricing · per 1M tokens · projected $/mo at 10M tokens
| Model | Input | Output | Context | Projected $/mo |
|---|---|---|---|---|
| $0.21 | $0.32 | 131K tokens (~66 books) | $2.41 | |
| $0.09 | $1.10 | 262K tokens (~131 books) | $3.43 |