The Known Good Updated 8 Oct 2026 Subscribe
DeepSeek: DeepSeek V3 vs DeepSeek: R1 ·DeepSeek vs DeepSeek

DeepSeek: DeepSeek V3 vs DeepSeek: R1 — head-to-head comparison

DeepSeek: DeepSeek V3 profile DeepSeek: R1 profile All comparisons
DeepSeek: DeepSeek V3 — DeepSeek
Quality
#43 of 65
62based on 4 of 7 evaluations
Known Good Index
Speed
#336 of 356
24.4
Output tokens per second
Price
#178 of 504
$0.45
Blended per 1M tokens
In $0.26 · Out $1.03
Cache hit price
#160 of 384
$0.050
per 1M tokens (−81%)
Context
#407 of 550
164k
tokens (~300 pages)
DeepSeek: R1 — DeepSeek
Quality
#22 of 65
75based on 4 of 7 evaluations
Known Good Index
Speed
#343 of 356
22.4
Output tokens per second
Price
#292 of 504
$1.15
Blended per 1M tokens
In $0.70 · Out $2.50
Cache hit price
unranked
—
per 1M tokens
Context
#517 of 550
64k
tokens (~98 pages)
Summary

On the Known Good Index, DeepSeek: DeepSeek V3 scores 62 and DeepSeek: R1 scores 75 — a 13-point lead for DeepSeek: R1. They share 10 published evaluations, with DeepSeek: DeepSeek V3 ahead on 1 and DeepSeek: R1 ahead on 9.

Blended pricing is $0.45 per 1M tokens for DeepSeek: DeepSeek V3 and $1.15 for DeepSeek: R1 — DeepSeek: DeepSeek V3 is about 2.6× cheaper.

Throughput is 24 tokens per second for DeepSeek: DeepSeek V3 and 22 for DeepSeek: R1, with time to first token at 2.6s and 1.2s. DeepSeek: DeepSeek V3 accepts 164k tokens of context against 64k.

DeepSeek: DeepSeek V3 is open weights under Open weights (unrestricted); DeepSeek: R1 is open weights under Open weights (unrestricted). 4 tracked hosts serve one or both, and 1 serve both.

Generated from tracked data on 9 Oct 2026 · how these comparisons are computed →

At a glance
MetricDeepSeek: DeepSeek V3DeepSeek: R1
Known Good Index 62 based on 4 of 7 evaluations 75 based on 4 of 7 evaluations
Blended $/1M $0.45 $1.15
Output tok/s 24 22
Context window 164k 64k
Providers 3 2
Jump to section

Quality

Known Good Index and the capability indices, side by side

Index scores

Each index is min-max normalised across tracked models and scaled 0–100 from published evaluations. We do not run these evaluations.

MetricDeepSeek: DeepSeek V3DeepSeek: R1
Known Good Index 62 75
Mathematics Index 57 74
Reasoning Index 62 77
Openness Index 60 60
The Known Good

Quality Breakdown

Every evaluation both models report a published score for

Evaluations both models report

Bar = DeepSeek: DeepSeek V3 · marker = DeepSeek: R1 · scores ingested from Epoch AI (CC-BY 4.0), LiveBench (Apache-2.0)

GPQA Diamond71 / 69
Mock AIME 2024–2549 / 53
MATH Level 565 / 93
LiveBench60 / 72
LiveBench Coding62 / 67
LiveBench Data Analysis61 / 70
LiveBench Instruction Following75 / 81
LiveBench Language47 / 49
LiveBench Mathematics61 / 81
LiveBench Reasoning57 / 83

Right-hand figures read DeepSeek: DeepSeek V3 / DeepSeek: R1. DeepSeek: DeepSeek V3 leads on 1 of 10, DeepSeek: R1 on 9.

The Known Good

Arena Elo

Human preference rating from blind pairwise votes

Arena ratings

Elo from blind pairwise votes, ingested from Arena. Higher is better.

MetricDeepSeek: DeepSeek V3DeepSeek: R1
Text Arena 1,332 1,373
Text Arena (style controlled) 1,358 1,398
The Known Good

Price & Cost

Published API pricing by token type

Published pricing

USD per 1M tokens. Blended is a 3:1 input:output mix. Live from OpenRouter, cross-checked against models.dev. Lower is better.

MetricDeepSeek: DeepSeek V3DeepSeek: R1
Input $/1M $0.26 $0.70
Output $/1M $1.03 $2.50
Blended $/1M $0.45 $1.15
Cache read $/1M $0.050 —
The Known Good
We publish blended price rather than cost-per-task. Cost-per-task requires token counts from an evaluation suite we do not run. Blended price is computed from published rates and is fully reproducible — see the pricing methodology.

Context Window

Maximum input accepted and maximum output emitted

Context and output limits

Provider-declared limits, from metadata via models.dev and OpenRouter. Higher is better.

MetricDeepSeek: DeepSeek V3DeepSeek: R1
Context window 164k 64k
Max output tokens 16k 16k
The Known Good

Speed & Latency

Mean throughput and time to first token across hosting providers, over successful measurements from the last 48 hours

Throughput and latency

Means across all tracked providers over successful measurements from the last 48 hours, derived from OpenRouter endpoint telemetry.

MetricDeepSeek: DeepSeek V3DeepSeek: R1
Output tokens/s 24 22
Time to first token 2.6s 1.2s
The Known Good

Providers

Which hosts serve each model, and at what price and speed

4 hosts · 1 serve both
Full provider benchmarks →
Provider DeepSeek: DeepSeek V3 blendedDeepSeek: DeepSeek V3 tok/s DeepSeek: R1 blendedDeepSeek: R1 tok/s
StreamLake $0.45 35 — —
Novita $0.62 17 $1.15 27
DeepInfra $0.46 10 — —
Azure · first party — — $2.60 —

Per-host pricing and telemetry from OpenRouter. A dash means that host does not serve that model.

The Known Good
Sources for this page
Pricing & providers — OpenRouter Metadata — models.dev Benchmarks — Epoch AI (CC-BY 4.0)Benchmarks — LiveBench (Apache-2.0)Elo — Arena Full attribution →