DeepSeek R1 0528 deepseek-r1-0528

deepseek-ai · deepseek reasoning flagship

unranked — no quality signals yet

params
671B (A37B)
arch
moe
context
160k
license
mit open weights
released
May 2025
reasoning
yes
train compute
4.0e+24 FLOP
downloads/30d
1.8M

# why ranked

Overall: provisional — score 49.5 ● low — a single signal; treat with caution

provisional: too few independent signals for a numbered position — the score is shown, but this model sorts after every ranked model.

signalweightinput (0–100)
aa_intelligence 0.50
arena_elo 0.30
bench_composite 0.20 49.5

missing signals are dropped and the remaining weights renormalized — never imputed.

Coding: provisional — score 90.9 ● low — a single signal; treat with caution

provisional: too few independent signals for a numbered position — the score is shown, but this model sorts after every ranked model.

signalweightinput (0–100)
aa_coding 0.25
aider_polyglot 0.25 90.9
arena_elo_code 0.15
swe_bench_verified 0.35

missing signals are dropped and the remaining weights renormalized — never imputed.

full methodology

# trend

collecting history — trends appear after a few daily runs.

# benchmarks

benchmarkscoresourcedate
Aider Polyglot 71.4 Aider Jun 2025
Aider-polyglot 0.7 / 1 LLM Stats
AIME 2024 0.9 / 1 LLM Stats
AIME 2025 0.9 / 1 LLM Stats
Browsecomp 0.1 / 1 LLM Stats
Browsecomp-zh 0.4 / 1 LLM Stats
Codeforces 0.6 / 3000 LLM Stats

# where to run

providerquantctx$/M in$/M out$/M blendtpsuptime
DeepInfrafp4 160k $0.50$2.15$0.91 100.0%
SiliconFlowfp8 160k $0.50$2.18$0.92 100.0%
StreamLakeunknown 125k $0.57$2.29$1.00 98.7%
Novitafp8 160k $0.70$2.50$1.15

blended = (3·input + output) / 4 per 1M tokens · cheapest first

source aliases
aa
deepseek-r1-0528
aider
DeepSeek R1 (0528)
arena
DeepSeek-R1-0528
epoch
DeepSeek-R1 (May 2025)
openrouter
deepseek/deepseek-r1-0528