Перейти к основному содержанию

Claude Sonnet 5.5 (Max, Default Fallback)

AnthropicClaudeProprietary

Описание

Claude Sonnet 5.5 is the second model in Anthropic's Claude 5.5 family and a faster, lower-cost complement to Claude Opus 5.5. It targets well-scoped everyday tasks, coding, and polished documents, slides, and spreadsheets, and runs 30%+ faster than Claude Sonnet 5 while typically using far fewer tokens at the same $2/$10 per million input/output prices. It accepts text and image input with text output, supports multilingual and vision workloads, tool use, and adaptive thinking (Claude Platform default effort high; Claude Code and apps default medium). Context window is 1M tokens with 128K max output on the sync Messages API; reliable knowledge and training-data cutoffs are June 2026. Cache reads are $0.10 and 5-minute cache writes are $2.50 per million tokens. Capability scores are self-reported from the Claude Sonnet 5.5 System Card (Table 8.1.A and §8) and the launch announcement. Available on the Claude API as `claude-sonnet-5-5`.

Дата выхода
2026-09-28
Параметры
—
Длина контекста
1.0M
Модальности
image, pdf, text

Радар способностей

55
general
70
coding
70
reasoning
51
scienceоцен.
60
agents
80
multimodal

При отсутствии специализированных научных бенчмарков Science оценивается по научным категориям LLM Stats или на основе рассуждений.

Рейтинги

Оценки бенчмарков (LLM Stats)

(LLM Stats (zeroeval))

Agents

AA-Briefcase v1.11811.00 / 3000Сам.
Program Bench79.7%Сам.
Toolathlon-Verified77.8%Сам.
OfficeQA76.9%Сам.
OfficeQA Pro65.6%Сам.
AutomationBench44.7%Сам.

Biology

BioMysteryBench89.2%Сам.
De novo protein binder design82.3%Сам.
LatchBio SpatialBench Verified72.5%Сам.
Biomedical image analysis72.2%Сам.
Protocols Troubleshooting67.3%Сам.
Protocols Understanding (V2)66.6%Сам.
Medicinal Chemistry (ADME)65.3%Сам.
Protein Design — Library Ranking54.8%Сам.
Protein Design — Sequence Generation51.0%Сам.
BioMysteryBench (Human Difficult)44.7%Сам.
Morphology-to-molecule matching25.0%Сам.

Code

BenchCAD (with Python tool)96.3%Сам.
BenchCAD74.7%Сам.
DeepSWE 1.171.0%Сам.
FrontierSWE V261.9%Сам.
CursorBench 4.055.5%Сам.

General

GDPval-AA 2.11844.00 / 3000Сам.
Global-MMLU92.1%Сам.

Healthcare

HealthBench Professional (raw)77.1%Сам.
HealthBench (raw)69.4%Сам.
HealthBench Professional69.2%Сам.
HealthBench65.4%Сам.
PhysicianBench63.2%Сам.

Language

MILU91.6%Сам.

Legal

Legal Agent Benchmark11.7%Сам.

Math

ArXivMath (with tools)95.2%Сам.
ArXivMath86.8%Сам.

Multimodal

OSWorld 2.1 (partial)80.1%Сам.
OSWorld 2.1 (strict)43.5%Сам.

Reasoning

SWE-bench Multilingual90.3%Сам.
SWE-Bench ProPrinceton NLP (2024)81.3%Сам.
Terminal-Bench 4.070.6%Сам.
Humanity's Last Exam (with tools)64.5%Сам.
FrontierCode 1.1 (Extended)64.4%Сам.
Terminal-Bench-Science 0.159.9%Сам.
Humanity's Last Exam (no tools, text-only)56.9%Сам.
SWE-Bench Multimodal54.3%Сам.
FrontierCode 1.152.1%Сам.

Science

LatchBio SingleCellBench59.1%Сам.

Vision

Chartography90.2%Сам.
Chartography (no tools)61.6%Сам.

Индексы оценки AA

(Artificial Analysis)
Engineering(Artificial Analysis)
58.5
Strategy And Ops
58.1
Finance And Accounting
55.8
Intelligence Index(Artificial Analysis)
54.5
Healthcare And Medical
54.2
Economics
54.1
Legal
52.2

Оценки категорий LLM Stats

(LLM Stats (zeroeval))
Productivity
100
Agents
100
Reasoning
100
General
90
Language
90
Math
70
Multimodal
70
Healthcare
70
Chemistry
70
Code
70
Vision
70
Science
60
Biology
60
Tool Calling
60
Legal
10

Цены

Цена ввода$2 / 1M токенов
Цена вывода$10 / 1M токенов
Смешанная цена (3:1)$4 / 1M токенов
Цена чтения кэша$0.1 / 1M токенов
Цена записи кэша$2.5 / 1M токенов

Скорость

Токенов/сек132.8
Задержка первого токена434.00s
Время до первого ответа434.00s

Рейтинг цен провайдеров

Рейтинг цен провайдеров

1 провайдеров

ПровайдерВводВывод
1AnthropicОсновной
$2
$10

Сравнение цен разных API-провайдеров для этой модели.

Внешние ссылки