Updated minutes ago
ReleasedJanuary 15, 2026Verified 3d ago · docs.anthropic.com
Anthropic

Claude Sonnet 4.6

Anthropic · moe · 180B parameters · 1,000,000 context

Quality
86.0
Deploy this model →Use via API →Calculate ROI →Compare with othersFine-tune this model

Parameters

180B

Context Window

977K tokens

Architecture

MoE

Best GPU

B200 NVL (pair)

Cheapest API

$15.00/M

Quality Score

86/100

Intelligence Brief

Claude Sonnet 4.6 is a 180B parameter Mixture-of-Experts (16 experts, 2 active) model from Anthropic, 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. On standardized benchmarks, it achieves MMLU 89, HumanEval 70, GSM8K 95. The most cost-effective API deployment is via anthropic at $15.00/M output tokens. For self-hosted inference, B200 NVL (pair) delivers optimal throughput at $19929/month.

Provider pricing

1 provider · canonical: anthropic
Provider Input $/M Output $/M ▲Notes
anthropiccanonical$3.00$15.00cheapest 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 Parameters180B
Active Parameters70B
Layers80
Hidden Dimension10,240
Attention Heads80
KV Heads10
Head Dimension128
Vocab Size200,000
Total Experts16
Active Experts2

Memory Requirements

BF16 Weights

360.0 GB

FP8 Weights

180.0 GB

INT4 Weights

90.0 GB

KV-Cache per Token204800 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

Claude Sonnet 4.6 is compatible with 14% of GPU configurations across 41 GPUs at 3 precision levels.

BF16 (Full)
FP8 (Half)
INT4 (Quarter)
Blackwell(7 GPUs)
B200 NVL (pair)360GB
✗
50%
75%
B300288GB
✗
38%
69%
B100 SXM192GB
✗
6%
53%
GB200 NVL72 (per GPU)192GB
✗
6%
53%
Hopper(7 GPUs)
H100 NVL 94GB (per GPU pair)188GB
✗
4%
52%
H200 SXM141GB
✗
✗
36%
H2096GB
✗
✗
6%
GH20096GB
✗
✗
6%
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

93/100

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

88/100

score

Throughput

280.0 tok/s

Latency (ITL)

3.6ms

Est. TTFT

1ms

Cost/Month

$17044

Cost/M Tokens

$23.16

Use this config →
B100 SXMoptimal

BF16 · 4 GPUs · tensorrt-llm

88/100

score

Throughput

280.0 tok/s

Latency (ITL)

3.6ms

Est. TTFT

1ms

Cost/Month

$17082

Cost/M Tokens

$23.21

Use this config →

Deployment Options

API

API Deployment

anthropic

$15.00/M

output tokens

Self-Hosted

Single GPU

Requires multi-GPU setup (180 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
anthropic$3.00$15.00
Cheapest

Cost Analysis

ProviderInput $/MOutput $/M~Monthly Cost
anthropicBest Value$3.00$15.00$90

Cost per 1,000 Requests

Short (500 tok)

$4.50

via anthropic

Medium (2K tok)

$18.00

via anthropic

Long (8K tok)

$54.00

via anthropic

Performance Estimates

Throughput by GPU

B200 NVL (pair)
280.0 tok/s
B200 SXM
280.0 tok/s
B100 SXM
280.0 tok/s

VRAM Breakdown (B200 NVL (pair), BF16)

Weights
Weights 180.0 GBKV-Cache 6.7 GBActivations 32.0 GBOverhead 9.0 GB

Quality Benchmarks

Top 10%
93th percentile across all models
MMLU
89.0
Above Average (86th pctile)
HumanEval
70.0
Above Average (80th pctile)
GSM8K
95.0
Above Average (83th pctile)
MT-Bench
89.0
Bottom 25% (0th pctile)

Capabilities

Features

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

Supported Frameworks

vllm

Supported Precisions

BF16 (default)

Where to Deploy Claude Sonnet 4.6

Similar Models

Claude 3 Opus

175B params · dense

Quality: 80

from $75.00/M

Smaller context, Lower quality, More expensiveCompare →

Claude 3.5 Sonnet

175B params · moe

Quality: 50

from $15.00/M

Smaller context, Lower qualityCompare →

Claude Opus 4

200B params · dense

Quality: 90

from $75.00/M

Smaller context, More expensiveCompare →

Claude Sonnet 4.5

150B params · moe

Quality: 86

from $15.00/M

Smaller contextCompare →

Claude 3 Sonnet

70B params · dense

Quality: 78

from $15.00/M

Smaller context, Lower quality, Smaller modelCompare →

Frequently Asked Questions

How much VRAM does Claude Sonnet 4.6 need for inference?

Claude Sonnet 4.6 requires approximately 360.0 GB of VRAM at BF16 precision, 180.0 GB at FP8, or 90.0 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 Claude Sonnet 4.6?

The top recommended GPU for Claude Sonnet 4.6 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: 93/100.

How much does Claude Sonnet 4.6 inference cost?

Claude Sonnet 4.6 API inference starts from $3.00/M input tokens and $15.00/M output tokens. Self-hosted inference costs depend on your GPU configuration — use our ROI calculator for a detailed breakdown.