modelbenchmark.io

DeepSeek-R1

DeepSeek · deepseek-r1

Compare

DeepSeek reasoning model for multi-step analysis, math, coding, and tools

Specification

deepseek listing

Context
66K
Max output
8K
Released
2025-08-05
Knowledge cutoff
2024-07
Retires
2026-08-13
Open weights
no
Input
text
Output
text

Price

US dollars per million tokens · deepseek list price

Input
$0.55
Output
$2.19
Cache read
$0.20
Cache write
$0.825

Quality

4 benchmarks · 2 sources

Composite
78th
percentile of 334 scored models
Rank
75 / 334
±0.05 sd
Evidence
4 × 2
3 or more benchmarks, from 2 or more sources
Effort range
one configuration only
coding58th1/3 bench
reasoning70th1/4 bench
math84th2/7 bench

Every score

one row per source and configuration — nothing averaged away

BenchmarkScoreConfigurationSourceRun
MATH level 593.0 ±0.7defaultEpoch AI2025-01-31
Aider polyglot56.9default · aider diffAider2025-01-20
GPQA diamond69.2 ±3.1defaultEpoch AI2025-05-26
OTIS Mock AIME 2024-202553.3 ±7.5defaultEpoch AI2025-02-26

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

Available from 46 hosts

HostIn $/MOut $/MCache rdCache wrContextOutputRetires
deepseekvendor$0.55$2.1966K8K
iflowcnfreefree128K32K
lambda_ai$0.20$0.60131K131K
hyperbolic
$0.40
$0.25
$0.40
$0.25
33K
131K
33K
131K
nano-gpt$0.40$1.70$0.20128K8K
nano-gpt · deepseek-ai$0.40$1.70$0.20164K33K
alibaba-cn · siliconflow$0.50$2.18164K33K
deepinfra · deepseek-ai$0.50$2.15$0.35164K64K
meganova · deepseek-ai$0.50$2.15164K64K
siliconflow · deepseek-ai$0.50$2.18164K164K
siliconflow-cn · deepseek-ai$0.50$2.18164K164K
siliconflow-cn · pro$0.50$2.18164K164K
submodel · deepseek-ai$0.50$2.1575K164K
drun · public$0.55$2.20131K32K
vercel_ai_gateway · vercel-ai-gateway$0.55$2.19128K8K
alibaba-cn$0.574$2.294131K16K
cortecs$0.652$2.57$0.163164K164K
tensorx$0.66$2.60$0.165$0.825164K8K
melious$0.6955$2.7821$0.1855164K33K
deepinfra
$0.70
$0.50
$2.40
$2.15
$0.40164K164K
edenai$0.70$2.40164K33K
jiekou$0.70$2.50164K33K
kilo$0.70$2.50$0.35
64K
164K
16K
33K
novita
$0.70
$4.00
$2.50
$4.00
$0.35
164K
64K
33K
16K
novita-ai$0.70$2.50$0.35164K33K
openrouter
$0.70
$0.50
$2.50
$2.15
$0.35
64K
164K
65K
16K
33K
8K
nebius$0.80$2.40
128K
164K
128K
164K
azure$1.35$5.40164K164K
azure_ai$1.35$5.40128K8K2026-08-13
azure-cognitive-services$1.35$5.40164K164K
merge-gateway$1.35$5.40164K41K
snowflake$1.35$5.40128K16K
vercel$1.35$5.40128K33K
wandb · wandb$1.35$5.40161K161K
io-net · deepseek-ai$2.00$8.75$1.00$4.00128K4K
abacus · deepseek-ai$3.00$7.00128K8K
crusoe · crusoe$3.00$7.00164K164K
fireworks_ai$3.00$8.00
128K
160K
20K
160K
huggingface · deepseek-ai
$3.00
$0.70
$5.00
$2.50
164K
64K
164K
33K
together_ai$3.00$7.00
128K
164K
20K2026-05-14
togetherai · deepseek-ai$3.00$7.00164K164K
replicate · replicate$3.75$10.0066K8K
sambanova · sambanova$5.00$7.0033K33K
anyapi1M384K
qiniu-ai128K32K
snowflake-cortex128K33K

Price history

append-only observations · a listing writes a row only when its price moves

DateHostField$/M
2026-09-02huggingface · deepseek-aiInput$0.70
2026-09-04huggingface · deepseek-aiInput$3.00
2026-09-02openrouterInput$0.70
2026-09-06openrouterInput$0.70
2026-09-02huggingface · deepseek-aiOutput$2.50
2026-09-04huggingface · deepseek-aiOutput$5.00
2026-09-02openrouterOutput$2.50
2026-09-06openrouterOutput$2.50

Other listings of this model

same model, different key

More from DeepSeek

most-hosted first