modelbenchmark.io

DeepSeek R1 Distill Qwen 32B

DeepSeek · deepseek-r1-distill-qwen-32b

Compare

Qwen instruction model for multilingual chat, reasoning, and tool use

Specification

most-agreed values

Context
33K
Max output
16K
Released
2025-01-01
Knowledge cutoff
2024-07
Retires
Open weights
no
Input
text
Output
text

Price

US dollars per million tokens · most-agreed

Input
$0.287
Output
$0.861

Rate limits

Requests per minute
300

Quality

3 benchmarks · 1 source

Composite
45th
percentile of 334 scored models
Rank
185 / 334
±0.10 sd
Evidence
3 × 1
one source, or fewer than 3 benchmarks
Effort range
one configuration only
reasoning40th2/4 bench
math60th1/7 bench

Every score

one row per source and configuration — nothing averaged away

BenchmarkScoreConfigurationSourceRun
OTIS Mock AIME 2024-202555.6 ±6.2defaultEpoch AI2026-08-28
GPQA diamond64.1 ±2.8defaultEpoch AI2026-08-28
Chess Puzzles1.0 ±1.0defaultEpoch AI2026-08-28

The leading figure for a benchmark is the median across its configurations, so one heroic high-effort run cannot set the number.

Available from 8 hosts

HostIn $/MOut $/MCache rdCache wrContextOutputRetires
nscale$0.15$0.15
deepinfra$0.27$0.27131K131K
alibaba-cn$0.287$0.86133K16K
novita$0.30$0.3064K32K
novita-ai$0.30$0.3064K32K
cloudflare$0.497$4.88180K80K
cloudflare-workers-ai · cf$0.497$4.88180K80K
fireworks_ai$0.90$0.90131K131K

More from DeepSeek

most-hosted first