Updated minutes ago
ReleasedFebruary 15, 2024Verified 3mo ago · ai.google.dev
Google

Gemini 1.5 Pro

Google · moe · 175B parameters · 2,097,152 context

Quality
80.0
Deploy this model →Use via API →Calculate ROI →Compare with othersFine-tune this model

Parameters

175B

Context Window

2048K tokens

Architecture

MoE

Best GPU

B200 NVL (pair)

Cheapest API

$5.00/M

Quality Score

80/100

Intelligence Brief

Gemini 1.5 Pro is a 175B parameter Mixture-of-Experts (16 experts, 2 active) model from Google, featuring Grouped Query Attention (GQA) with 72 layers and 8,192 hidden dimensions. With a 2,097,152 token context window, it supports tools, vision, structured output, code, math, multilingual. On standardized benchmarks, it achieves MMLU 86.5, HumanEval 65, GSM8K 92. The most cost-effective API deployment is via google at $5.00/M output tokens. For self-hosted inference, B200 NVL (pair) delivers optimal throughput at $19929/month.

Provider pricing

1 provider · canonical: google
Provider Input $/M Output $/M ▲Notes
googlecanonical$1.25$5.00cheapest input · cheapest output

Prices update via the nightly pricing cron + admin approvals at /admin/ingest-queue. The leaderboard's Input/Output cells show the canonical rate above; this table shows the full spread.

Recent changes

Loading…

Related models

5 suggestions

Picks: same family first, then same vendor within ±2× params, then top tag-overlap matches. Price shown is the cheapest Output $/M across providers — the row's page shows the canonical anchor.

Architecture Details

TypeMOE
Total Parameters175B
Active Parameters40B
Layers72
Hidden Dimension8,192
Attention Heads64
KV Heads8
Head Dimension128
Vocab Size256,000
Total Experts16
Active Experts2

Memory Requirements

BF16 Weights

350.0 GB

FP8 Weights

175.0 GB

INT4 Weights

87.5 GB

KV-Cache per Token147456 bytes
Activation Estimate4.00 GB

Fits on (single GPU) — most practical first

BEST FITInstinct MI250X128GBINT4Gaudi 3128GBINT4Gaudi 3 HL-325L128GBINT4H200 SXM141GBINT4B200 SXM180GBINT4H100 NVL 94GB (per GPU pair)188GBINT4B100 SXM192GBINT4GB200 NVL72 (per GPU)192GBINT4GB300 NVL72 (per GPU)192GBINT4Instinct MI300X192GBINT4Groq LPU230GBFP8Instinct MI325X256GBFP8

GPU Compatibility Matrix

Gemini 1.5 Pro is compatible with 15% of GPU configurations across 41 GPUs at 3 precision levels.

BF16 (Full)
FP8 (Half)
INT4 (Quarter)
Blackwell(7 GPUs)
B200 NVL (pair)360GB
3%
51%
76%
B300288GB
✗
39%
70%
B100 SXM192GB
✗
9%
54%
GB200 NVL72 (per GPU)192GB
✗
9%
54%
Hopper(7 GPUs)
H100 NVL 94GB (per GPU pair)188GB
✗
7%
53%
H200 SXM141GB
✗
✗
38%
H2096GB
✗
✗
9%
GH20096GB
✗
✗
9%
Ada Lovelace(11 GPUs)
L40S48GB
✗
✗
✗
L4048GB
✗
✗
✗
RTX 6000 Ada48GB
✗
✗
✗
L2048GB
✗
✗
✗
Ampere(16 GPUs)
A100 80GB SXM80GB
✗
✗
✗
A100 80GB PCIe80GB
✗
✗
✗
A1664GB
✗
✗
✗
RTX A600048GB
✗
✗
✗
Legend:No fitVery tightTightModerateGoodExcellent

GPU Recommendations

B200 NVL (pair)optimal

BF16 · 2 GPUs · tensorrt-llm

98/100

score

Throughput

280.0 tok/s

Latency (ITL)

3.6ms

Est. TTFT

1ms

Cost/Month

$19929

Cost/M Tokens

$27.08

Use this config →
B200 SXMoptimal

BF16 · 4 GPUs · tensorrt-llm

93/100

score

Throughput

280.0 tok/s

Latency (ITL)

3.6ms

Est. TTFT

1ms

Cost/Month

$17044

Cost/M Tokens

$23.16

Use this config →
H200 SXMoptimal

BF16 · 4 GPUs · tensorrt-llm

90/100

score

Throughput

280.0 tok/s

Latency (ITL)

3.6ms

Est. TTFT

1ms

Cost/Month

$10211

Cost/M Tokens

$13.88

Use this config →

Deployment Options

API

API Deployment

google

$5.00/M

output tokens

Self-Hosted

Single GPU

Requires multi-GPU setup (175 GB VRAM needed)

Scale

Multi-GPU

B200 NVL (pair) x2

280.0 tok/s

TP· $19929/mo

Open Full Calculator →

API Pricing Comparison

ProviderInput $/MOutput $/MBadges
google$1.25$5.00
Cheapest

Cost Analysis

ProviderInput $/MOutput $/M~Monthly Cost
googleBest Value$1.25$5.00$31

Cost per 1,000 Requests

Short (500 tok)

$1.63

via google

Medium (2K tok)

$6.50

via google

Long (8K tok)

$20.00

via google

Performance Estimates

Throughput by GPU

B200 NVL (pair)
280.0 tok/s
B200 SXM
280.0 tok/s
H200 SXM
280.0 tok/s

VRAM Breakdown (B200 NVL (pair), BF16)

Weights
Weights 175.0 GBKV-Cache 4.8 GBActivations 32.0 GBOverhead 8.8 GB

Quality Benchmarks

Above Average
85th percentile across all models
MMLU
86.5
Above Average (77th pctile)
HumanEval
65.0
Average (74th pctile)
GSM8K
92.0
Average (72th pctile)
MT-Bench
87.0
Bottom 25% (0th pctile)

Capabilities

Features

✓ Tool Use✓ Vision✓ Code✓ Math✗ Reasoning✓ Multilingual✓ Structured Output

Supported Frameworks

Supported Precisions

BF16 (default)

Where to Deploy Gemini 1.5 Pro

Similar Models

Claude 3 Opus

175B params · dense

Quality: 80

from $75.00/M

Smaller context, More expensiveCompare →

Claude 3.5 Sonnet

175B params · moe

Quality: 50

from $15.00/M

Smaller context, Lower quality, More expensiveCompare →

Claude Sonnet 4.6

180B params · moe

Quality: 86

from $15.00/M

Smaller context, Higher quality, More expensiveCompare →

Falcon 180B

180B params · dense

Quality: 60

from $2.40/M

Smaller context, Lower quality, CheaperCompare →

Yi-Lightning

200B params · moe

Quality: 50

from $0.99/M

Smaller context, Lower quality, CheaperCompare →

Frequently Asked Questions

How much VRAM does Gemini 1.5 Pro need for inference?

Gemini 1.5 Pro requires approximately 350.0 GB of VRAM at BF16 precision, 175.0 GB at FP8, or 87.5 GB at INT4 quantization. Additional VRAM is needed for KV-cache (147456 bytes per token) and activations (~4.00 GB).

What is the best GPU for Gemini 1.5 Pro?

The top recommended GPU for Gemini 1.5 Pro is the B200 NVL (pair) (x2) using BF16 precision. It achieves approximately 280.0 tokens/sec at an estimated cost of $19929/month ($27.08/M tokens). Score: 98/100.

How much does Gemini 1.5 Pro inference cost?

Gemini 1.5 Pro API inference starts from $1.25/M input tokens and $5.00/M output tokens. Self-hosted inference costs depend on your GPU configuration — use our ROI calculator for a detailed breakdown.