The Known Good Updated 25 Jul 2026
DeepSeek: DeepSeek V3 0324 vs Google: Gemini 2.5 Pro ·DeepSeek vs Google

DeepSeek: DeepSeek V3 0324 vs Google: Gemini 2.5 Pro — head-to-head comparison

DeepSeek: DeepSeek V3 0324 profile Google: Gemini 2.5 Pro profile All comparisons
DeepSeek: DeepSeek V3 0324 — DeepSeek
Quality
#35 of 61
60based on 3 of 6 evaluations
Known Good Index
Speed
#256 of 316
26.1
Output tokens per second
Price
#133 of 340
$0.48
Blended per 1M tokens
In $0.27 · Out $1.12
Cache hit price
#118 of 206
$0.14
per 1M tokens (−50%)
Context
#224 of 345
164k
tokens (~300 pages)
Google: Gemini 2.5 Pro — Google
Quality
#29 of 61
64based on 4 of 6 evaluations
Known Good Index
Speed
#53 of 316
96.4
Output tokens per second
Price
#262 of 340
$3.44
Blended per 1M tokens
In $1.25 · Out $10.00
Cache hit price
#106 of 206
$0.12
per 1M tokens (−90%)
Context
#29 of 345
1.05M
tokens (~1,600 pages)
Summary

On the Known Good Index, DeepSeek: DeepSeek V3 0324 scores 60 and Google: Gemini 2.5 Pro scores 64 — a 5-point lead for Google: Gemini 2.5 Pro. They share 2 published evaluations, with DeepSeek: DeepSeek V3 0324 ahead on 0 and Google: Gemini 2.5 Pro ahead on 2.

Blended pricing is $0.48 per 1M tokens for DeepSeek: DeepSeek V3 0324 and $3.44 for Google: Gemini 2.5 Pro — DeepSeek: DeepSeek V3 0324 is about 7.1× cheaper.

Throughput is 26 tokens per second for DeepSeek: DeepSeek V3 0324 and 96 for Google: Gemini 2.5 Pro, with time to first token at 1.1s and 5.4s. DeepSeek: DeepSeek V3 0324 accepts 164k tokens of context against 1.05M.

DeepSeek: DeepSeek V3 0324 is open weights; Google: Gemini 2.5 Pro is proprietary. 6 tracked hosts serve one or both.

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

At a glance
MetricDeepSeek: DeepSeek V3 0324Google: Gemini 2.5 Pro
Known Good Index 60 based on 3 of 6 evaluations 64 based on 4 of 6 evaluations
Blended $/1M $0.48 $3.44
Output tok/s 26 96
Context window 164k 1.05M
Providers 4 2
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: DeepSeek V3 0324Google: Gemini 2.5 Pro
Known Good Index 60 64
Coding Index — 42
Mathematics Index 57 84
Science Index 67 89
Reasoning Index 65 —
Agentic Index — 42
Openness Index 100 5
The Known Good

Quality Breakdown

Every evaluation both models report a published score for

Evaluations both models report

Bar = DeepSeek: DeepSeek V3 0324 · marker = Google: Gemini 2.5 Pro · scores ingested from Epoch AI (CC-BY 4.0)

GPQA Diamond68 / 85
MATH / AIME38 / 84

Right-hand figures read DeepSeek: DeepSeek V3 0324 / Google: Gemini 2.5 Pro. DeepSeek: DeepSeek V3 0324 leads on 0 of 2, Google: Gemini 2.5 Pro 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: DeepSeek V3 0324Google: Gemini 2.5 Pro
Text Arena 1,375 1,457
Text Arena (style controlled) 1,396 1,446
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: DeepSeek V3 0324Google: Gemini 2.5 Pro
Input $/1M $0.27 $1.25
Output $/1M $1.12 $10.00
Blended $/1M $0.48 $3.44
Cache read $/1M $0.135 $0.125
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 V3 0324Google: Gemini 2.5 Pro
Context window 164k 1.05M
Max output tokens 66k 66k
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: DeepSeek V3 0324Google: Gemini 2.5 Pro
Output tokens/s 26 96
Time to first token 1.1s 5.4s
The Known Good

Providers

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

6 hosts
Full provider benchmarks →
Provider DeepSeek: DeepSeek V3 0324 blendedDeepSeek: DeepSeek V3 0324 tok/s Google: Gemini 2.5 Pro blendedGoogle: Gemini 2.5 Pro tok/s
Crusoe $0.75 37 — —
Novita $0.48 30 — —
DeepInfra $0.41 25 — —
SiliconFlow $0.44 21 — —
Google AI Studio — — $1.72 118
Google · first party — — $1.72 99

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 →