Updated minutes ago
ReleasedApril 24, 2026Verified 3d ago · openrouter.ai
OpenAI

GPT-5.5

OpenAI · moe · 700B parameters · 1,000,000 context

Quality
50.0
Deploy this model →Use via API →Calculate ROI →Compare with othersFine-tune this model

Parameters

700B

Context Window

977K tokens

Architecture

MoE

Best GPU

B200 NVL (pair)

Cheapest API

$30.00/M

Intelligence Brief

GPT-5.5 is a 700B parameter Mixture-of-Experts (16 experts, 2 active) model from OpenAI, featuring Grouped Query Attention (GQA) with 80 layers and 10,240 hidden dimensions. With a 1,000,000 token context window, it supports tools, vision, structured output, code, math, multilingual, reasoning. The most cost-effective API deployment is via openai at $30.00/M output tokens. For self-hosted inference, B200 NVL (pair) delivers optimal throughput at $79716/month.

Provider pricing

1 provider · canonical: openai
Provider Input $/M Output $/M ▲Notes
openaicanonical$5.00$30.00cheapest 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 Parameters700B
Active Parameters110B
Layers80
Hidden Dimension10,240
Attention Heads80
KV Heads10
Head Dimension128
Vocab Size200,000
Total Experts16
Active Experts2

Memory Requirements

BF16 Weights

1400.0 GB

FP8 Weights

700.0 GB

INT4 Weights

350.0 GB

KV-Cache per Token204800 bytes
Activation Estimate4.00 GB

Fits on (multi-GPU with Tensor Parallelism)

RECOMMENDEDx2Groq LPU230GB eachINT4x2Instinct MI325X256GB eachINT4x2B300288GB eachINT4x2B200 NVL (pair)360GB eachINT4x3H200 SXM141GB eachINT4x3B200 SXM180GB eachINT4x3H100 NVL 94GB (per GPU pair)188GB eachINT4x3B100 SXM192GB eachINT4x3GB200 NVL72 (per GPU)192GB eachINT4x3GB300 NVL72 (per GPU)192GB eachINT4x3Instinct MI300X192GB eachINT4x4Instinct MI250X128GB eachINT4

Multi-GPU configurations use Tensor Parallelism (TP) to split model layers across GPUs. Requires NVLink or NVSwitch interconnect for optimal performance.

This model requires multi-GPU deployment. Minimum: 2x Groq LPU (230GB each) with Tensor Parallelism.

GPU Compatibility Matrix

GPT-5.5 is compatible with 1% of GPU configurations across 41 GPUs at 3 precision levels.

BF16 (Full)
FP8 (Half)
INT4 (Quarter)
Blackwell(7 GPUs)
B200 NVL (pair)360GB
✗
✗
3%
B300288GB
✗
✗
✗
B100 SXM192GB
✗
✗
✗
GB200 NVL72 (per GPU)192GB
✗
✗
✗
Hopper(7 GPUs)
H100 NVL 94GB (per GPU pair)188GB
✗
✗
✗
H200 SXM141GB
✗
✗
✗
H2096GB
✗
✗
✗
GH20096GB
✗
✗
✗
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)good

BF16 · 8 GPUs · tensorrt-llm

73/100

score

Throughput

140.0 tok/s

Latency (ITL)

7.1ms

Est. TTFT

1ms

Cost/Month

$79716

Cost/M Tokens

$216.67

Use this config →
Instinct MI325Xgood

BF16 · 8 GPUs · vllm

65/100

score

Throughput

140.0 tok/s

Latency (ITL)

7.1ms

Est. TTFT

1ms

Cost/Month

$18904

Cost/M Tokens

$51.38

Use this config →
B200 SXMgood

BF16 · 16 GPUs · tensorrt-llm

63/100

score

Throughput

140.0 tok/s

Latency (ITL)

7.1ms

Est. TTFT

1ms

Cost/Month

$68176

Cost/M Tokens

$185.30

Use this config →

Deployment Options

API

API Deployment

openai

$30.00/M

output tokens

Self-Hosted

Single GPU

Requires multi-GPU setup (700 GB VRAM needed)

Scale

Multi-GPU

B200 NVL (pair) x8

140.0 tok/s

TP· $79716/mo

Open Full Calculator →

API Pricing Comparison

ProviderInput $/MOutput $/MBadges
openai$5.00$30.00
Cheapest

Cost Analysis

ProviderInput $/MOutput $/M~Monthly Cost
openaiBest Value$5.00$30.00$175

Cost per 1,000 Requests

Short (500 tok)

$8.50

via openai

Medium (2K tok)

$34.00

via openai

Long (8K tok)

$100.00

via openai

Performance Estimates

Throughput by GPU

B200 NVL (pair)
140.0 tok/s
Instinct MI325X
140.0 tok/s
B200 SXM
140.0 tok/s

VRAM Breakdown (B200 NVL (pair), BF16)

Weights
Weights 175.0 GBKV-Cache 6.7 GBActivations 32.0 GBOverhead 8.8 GB

Capabilities

Features

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

Supported Frameworks

vllm

Supported Precisions

BF16 (default)

Where to Deploy GPT-5.5

Similar Models

GPT-5

500B params · moe

Quality: 50

from $10.00/M

Smaller context, CheaperCompare →

GPT-4.5 Preview

1500B params · moe

Quality: 93

from $150.00/M

Smaller context, Higher quality, More expensive, Larger modelCompare →

DeepSeek V3-0324

685B params · moe

Quality: 81

from $0.00/M

Smaller context, Higher quality, CheaperCompare →

DeepSeek R1

671B params · moe

Quality: 88

from $2.00/M

Smaller context, Higher quality, CheaperCompare →

DeepSeek V3

671B params · moe

Quality: 81

from $0.42/M

Smaller context, Higher quality, CheaperCompare →

Frequently Asked Questions

How much VRAM does GPT-5.5 need for inference?

GPT-5.5 requires approximately 1400.0 GB of VRAM at BF16 precision, 700.0 GB at FP8, or 350.0 GB at INT4 quantization. Additional VRAM is needed for KV-cache (204800 bytes per token) and activations (~4.00 GB).

What is the best GPU for GPT-5.5?

The top recommended GPU for GPT-5.5 is the B200 NVL (pair) (x8) using BF16 precision. It achieves approximately 140.0 tokens/sec at an estimated cost of $79716/month ($216.67/M tokens). Score: 73/100.

How much does GPT-5.5 inference cost?

GPT-5.5 API inference starts from $5.00/M input tokens and $30.00/M output tokens. Self-hosted inference costs depend on your GPU configuration — use our ROI calculator for a detailed breakdown.