The Known Good Updated 25 Jul 2026
DeepSeek: R1 vs grok-4-0709 ·DeepSeek vs xAI

DeepSeek: R1 vs grok-4-0709 — head-to-head comparison

DeepSeek: R1 profile grok-4-0709 profile All comparisons
DeepSeek: R1 — DeepSeek
Quality
#24 of 61
68based on 3 of 6 evaluations
Known Good Index
Speed
#165 of 316
49.3
Output tokens per second
Price
#209 of 340
$1.15
Blended per 1M tokens
In $0.70 · Out $2.50
Cache hit price
unranked
—
per 1M tokens
Context
#219 of 345
164k
tokens (~300 pages)
grok-4-0709 — xAI
Quality
#27 of 61
67based on 3 of 6 evaluations
Known Good Index
Speed
unranked
—
Output tokens per second
Price
unranked
—
Blended per 1M tokens
Cache hit price
unranked
—
per 1M tokens
Context
unranked
—
tokens
Summary

On the Known Good Index, DeepSeek: R1 scores 68 and grok-4-0709 scores 67, which puts them level. They share 2 published evaluations, with DeepSeek: R1 ahead on 0 and grok-4-0709 ahead on 2.

Published pricing is tracked for DeepSeek: R1 only, at $1.15 per 1M tokens blended.

DeepSeek: R1 is open weights under Open weights (unrestricted); grok-4-0709 is proprietary. 2 tracked hosts serve one or both.

Generated from tracked data on 25 Jul 2026 · how these comparisons are computed →

At a glance
MetricDeepSeek: R1grok-4-0709
Known Good Index 68 based on 3 of 6 evaluations 67 based on 3 of 6 evaluations
Blended $/1M $1.15 —
Output tok/s 49 —
Context window 164k —
Providers 2 0
Jump to section

Quality

Known Good Index and the capability indices, side by side

IndicesComponents
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: R1grok-4-0709
Known Good Index 68 67
Mathematics Index 74 84
Science Index 69 91
Reasoning Index 77 —
Agentic Index — 27
Openness Index 60 5
The Known Good

Quality Breakdown

Every evaluation both models report a published score for

Evaluations both models report

Bar = DeepSeek: R1 · marker = grok-4-0709 · scores ingested from Epoch AI (CC-BY 4.0)

GPQA Diamond69 / 87
MATH / AIME53 / 84

Right-hand figures read DeepSeek: R1 / grok-4-0709. DeepSeek: R1 leads on 0 of 2, grok-4-0709 on 2.

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: R1grok-4-0709
Text Arena 1,373 1,410
Text Arena (style controlled) 1,398 1,410
The Known Good

Price & Cost

Published API pricing by token type

Per token typeCache discount
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: R1grok-4-0709
Input $/1M $0.70 —
Output $/1M $2.50 —
Blended $/1M $1.15 —
Cache read $/1M — —
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: R1grok-4-0709
Context window 164k —
Max output tokens 16k —
The Known Good

Speed & Latency

Median throughput and time to first token across hosting providers

Output speedTime to first token
Throughput and latency

Medians across all tracked providers, derived from OpenRouter endpoint telemetry.

MetricDeepSeek: R1grok-4-0709
Output tokens/s 49 —
Time to first token 1.1s —
The Known Good

Providers

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

2 hosts
Full provider benchmarks →
Provider DeepSeek: R1 blendedDeepSeek: R1 tok/s grok-4-0709 blendedgrok-4-0709 tok/s
Azure · first party $2.60 77 — —
Novita $1.15 22 — —

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)Elo — Arena Full attribution →