Why is this comparison indexable?

It has 6 verifiable comparison points, and both models have pricing or benchmark data.

Are missing metrics invented?

No. Metrics without LMSpeed data are omitted from this report.

Back to models

Data points: 138

Model compare

Claude Sonnet 4.6 vs GPT-5.4

The readout for Claude Sonnet 4.6 and GPT-5.4, before the detailed comparison sheet.

Model A

Claude Sonnet 4.6

Anthropic

Contender

Model B

GPT-5.4

OpenAI

Leading

Key Takeaways

Weighted outcome: GPT-5.4. Benchmark capability categories carry 80%, while price, API performance, and availability carry 20%.

Decision read

GPT-5.4

GPT-5.4 has the higher weighted result; Model A / B score 19.4 to 80.6.

Evidence depth

138 data points

Includes 37 benchmark rows, 0 audit samples, and 10 provider examples.

Selection signal

Start with GPT-5.4

The charts below split 47 high-signal samples across speed, scores, and audit health.

Change comparison

Switch either side of this report to compare another model with the same LMSpeed data pipeline.

Model AModel B

Comparison sheet

This report only uses LMSpeed data for Claude Sonnet 4.6 and GPT-5.4: pricing, speed aggregates, third-party benchmark scores, and shared provider samples.

Model compare	Claude Sonnet 4.6	GPT-5.4
Overall leader	Contender	Leading
Weighted overall score	19.4 pts	80.6 pts
Benchmark category leads	1 categories	6 categories
Operational advantages	Average speed, First-token latency	Cheapest input price, Free providers, Provider coverage
Context window	1M tokens	1.1M tokens
Max output	128K tokens	128K tokens
Modalities	Input TextImageFile Output Text	Input TextImageFile Output Text
Features	Text inputImage inputFile inputText outputTool callingStructured outputsJSON modeReasoning	Text inputImage inputFile inputText outputTool callingStructured outputsJSON modeReasoning

The overall result weights benchmark capability categories at 80% and price, API speed/latency, and availability at 20%. Recent test volume does not affect the winner, and missing benchmark categories are excluded.

Model metadata

Model compare	Claude Sonnet 4.6	GPT-5.4
Developer	Anthropic	OpenAI
Released	Feb 2026	Mar 2026
Parameters	No data	No data
Tokenizer	Claude	GPT
Knowledge cutoff	No data	No data
OpenRouter ID	anthropic/claude-sonnet-4.6	openai/gpt-5.4
References	No data	No data

When to choose each model

This report only uses LMSpeed data for Claude Sonnet 4.6 and GPT-5.4: pricing, speed aggregates, third-party benchmark scores, and shared provider samples.

Claude Sonnet 4.6

Claude Sonnet 4.6 is stronger in benchmark categories (Knowledge) and operational dimensions (Average speed, First-token latency).

GPT-5.4

GPT-5.4 is stronger in benchmark categories (Agents, Coding, Reasoning, Math, Multimodal) and operational dimensions (Cheapest input price, Free providers, Provider coverage).

Benchmark score comparison

Third-party benchmark profile synced into LMSpeed; only metrics available for both models are shown.

Category performance

Compare benchmark category scores on a 0-100 scale. Select a category to inspect the gap.

Model A coverage: 7 / 8
Model B coverage: 7 / 8
Shared: 7 shared categories

Avg. score

Claude Sonnet 4.6

52.9

Avg. score

GPT-5.4

56.6

Agents

GPT-5.4 leads by 4.4

Claude Sonnet 4.651.9

GPT-5.456.3

Coding

GPT-5.4 leads by 6.2

Claude Sonnet 4.656

GPT-5.462.2

Reasoning

GPT-5.4 leads by 5.6

Claude Sonnet 4.650.1

GPT-5.455.7

Knowledge

Claude Sonnet 4.6 leads by 12.1

Claude Sonnet 4.670.6

GPT-5.458.5

Math

GPT-5.4 leads by 4.2

Claude Sonnet 4.652.6

GPT-5.456.8

Multilingual

No data

Claude Sonnet 4.6-

GPT-5.4-

Multimodal

GPT-5.4 leads by 7.1

Claude Sonnet 4.645.4

GPT-5.452.5

Instruction following

GPT-5.4 leads by 10.2

Claude Sonnet 4.643.8

GPT-5.454

Professional benchmark details

Metric-level scores with benchmark source, rank depth, confidence, error, and evaluation date where available.

Group

Aggregatereported

Provider	Claude Sonnet 4.6	GPT-5.4
天宫造物75 tests	Claude Sonnet 4.6 speed / latency 40 tok/s / 3145ms input / output No data	GPT-5.4 speed / latency 50 tok/s / 7305ms input / output No data
巨量API51 tests	Claude Sonnet 4.6 speed / latency 39 tok/s / 5723ms input / output No data	GPT-5.4 speed / latency 43 tok/s / 12778ms input / output No data
UoCode30 tests	Claude Sonnet 4.6 speed / latency 32 tok/s / 6074ms input / output No data	GPT-5.4 speed / latency 31 tok/s / 2364ms input / output No data
Sub2API20 tests	Claude Sonnet 4.6 speed / latency N/A / N/A input / output No data	GPT-5.4 speed / latency 34 tok/s / 1748ms input / output No data
6345ywz API15 tests	Claude Sonnet 4.6 speed / latency N/A / N/A input / output No data	GPT-5.4 speed / latency 41 tok/s / 6581ms input / output No data
180txt API	Claude Sonnet 4.6 claude-sonnet-4-6 speed / latency No data input / output $0.450/M/$2.25/M	GPT-5.4 gpt-5.4-2026-03-05 speed / latency No data input / output $0/M/$0/M
APIMart	Claude Sonnet 4.6 claude-sonnet-4-6 speed / latency No data input / output $2.30/M/$11.51/M	GPT-5.4 gpt-5.4 speed / latency No data input / output $0/M/$0/M
兔子API	Claude Sonnet 4.6 claude-sonnet-4-6 speed / latency No data input / output $0/request	GPT-5.4 gpt-5.4-2026-03-05 speed / latency No data input / output $0/request
熊猫 API	Claude Sonnet 4.6 熊猫-特供-按量-满血逆-claude-sonnet-4.6 speed / latency No data input / output $0.0078/M/$0.039/M	GPT-5.4 熊猫-按量-gpt-5.4 speed / latency No data input / output $0.0054/M/$0.043/M
KFCV50	Claude Sonnet 4.6 claude-sonnet-4-6 speed / latency No data input / output $0.010/request	GPT-5.4 gpt-5.4 speed / latency No data input / output $0.014/request

Comparison sheet

Model metadata

When to choose each model

Benchmark score comparison

Category performance

Agents

Coding

Reasoning

Knowledge

Math

Multilingual

Multimodal

Instruction following

Professional benchmark details

BenchLM overall score

Input price

Blended price

Output price

CritPt

AA-LCR

BenchLM Reasoning score

HLE

GPQA

React Native Evals

Vibe Code Bench

EEBench

AA-SciCode

BenchLM Coding score

FrontierMath v2 (Tiers 1-3)

FrontierMath v2 (Tier 4)

BenchLM Math score

AA-Omniscience Hallucination Rate

AA-HLE

AA-Omniscience Accuracy

AA-GPQA Diamond

BenchLM Knowledge score

Artificial Analysis Intelligence Index

CharXiv

Design Arena Website

AA-MMMU-Pro

BenchLM Multimodal Grounded score

BenchLM Instruction Following score

AA-IFBench

τ²-bench results

Gert Labs

Claw-Eval

CyberGym

JobBench

OSWorld-Verified

Terminal-Bench 2.0

BenchLM Agentic score

API audit comparison

Provider examples

FAQ

Related compare reports