Compare AI models

Specifications
Parameters
671B27B/35B/122B
Context window
128k256k
API pricing
Cheapest input
$0.259SiliconFlow—
Cheapest output
$0.38DeepInfra—
$71.00DeepInfra · fp4—
Benchmarks
SWE-Bench Verified
73.1%72%
LiveCodeBench
83.3%78.9%
Terminal-Bench 2.0
46.4%49.4%
BrowseComp
51.4%63.8%
GPQA Diamond
82.4%86.6%
MMLU-Pro
85%86.7%
Overview
CompanyDeepSeekQwen
Release dateDec 1 2025Feb 24 2026
AccessOpen WeightOpen Weight
Model detailsView modelView model

Frequently asked questions

Qwen3.5 Medium leads DeepSeek-V3.2 on 4 of the 6 benchmarks they both report. DeepSeek-V3.2 shipped 85 days before Qwen3.5 Medium, so benchmark comparisons should account for the intervening progress.

DeepSeek-V3.2 has 671B parameters, while Qwen3.5 Medium has 27B/35B/122B. Context windows are 128k (DeepSeek-V3.2) vs 256k (Qwen3.5 Medium).

On SWE-Bench Verified, DeepSeek-V3.2 leads at 73.1% vs Qwen3.5 Medium at 72%. On LiveCodeBench, DeepSeek-V3.2 leads at 83.3% vs Qwen3.5 Medium at 78.9%. On Terminal-Bench 2.0, Qwen3.5 Medium leads at 49.4% vs DeepSeek-V3.2 at 46.4%. On BrowseComp, Qwen3.5 Medium leads at 63.8% vs DeepSeek-V3.2 at 51.4%. On GPQA Diamond, Qwen3.5 Medium leads at 86.6% vs DeepSeek-V3.2 at 82.4%. On MMLU-Pro, Qwen3.5 Medium leads at 86.7% vs DeepSeek-V3.2 at 85%.