Skip to content
Tier A — Frontier
Runs in:GlobalMade in:United States
Anthropic

Claude Sonnet 5.5

Tier A — Frontier · 1M tokens

Tokonomix Editorial Team·Reviewed by Mes Kalkan··
Section 01

Speed analysis

Latency measured across all benchmark runs. P50 (median) and P95 (95th percentile) give a realistic picture of response speed under normal and peak load.

P50 latency (median)P95 latency119 runs
599276449307095926010-2311-22ms
Section 02

Pricing

What you pay per million tokens when you use this model on Tokonomix, plus an estimate for a typical conversation.

💰
API rates — Claude Sonnet 5.5
$2.60 per 1M input tokens
$13.00 per 1M output tokens
≈ $0.0042 per typical conversation (800 tokens)
Input vs output price (per 1M tokens)
per 1M input tokens$2.60
per 1M output tokens$13.00
Section 03

Tokens per second

Throughput in tokens per second, derived from measured P50 latency. Higher is better; fluctuations track provider-side load.

Throughput (tokens / s)26 / avg 173
33119

Estimated from P50 latency × 200 output tokens — the absolute number depends on this assumption; the trend is what matters.

Section 04

Capabilities

toolssource: seed-newest-generation-2026-09-30visionjson modepdf inputreasoningjson schemaprompt cachingmax output tokens: 128000
Section 05

Availability

Availability

How often this model answers when we call it — measured across real API requests and live tests over the last 30 days. This is separate from quality: these numbers only tell you whether the model responds, not how good the answer is.

Last 7 days

100.0%

n=15

Last 30 days

100.0%

n=255

Median response time

7,919ms

n=255

Based on 612 measurements over the last 30 days.

Technical details

Only live API calls and live-test requests count — internal probes and benchmark runs are excluded.

Calls with a custom API key (BYOK) are excluded: those failures are key-specific, not a sign of model downtime.

Failed calls are NOT included in quality scores — quality is measured on successful responses only. Availability and quality are independent signals.

Median response time (p50) across successful calls with a recorded duration. Outliers (very slow or very fast calls) pull the median less than the average.

Total calls (30d)

255

OK responses (30d)

255

Total calls (7d)

15

OK responses (7d)

15

Section 06

Tokonomix benchmark verdicts

● 2026-11-22

Claude Sonnet 3.5 upgraded to 5.5 with expanded multimodal capabilities

Claude Sonnet has been upgraded from version 3.5 to 5.5, introducing significant new capabilities while maintaining the same benchmark performance characteristics. The model now supports seven distinct features: tools, vision, JSON mode, PDF input, reasoning, JSON schema, and prompt caching. These additions transform Claude Sonnet from a text-only model into a comprehensive multimodal system capable of processing images and documents, structured output generation, and enhanced reasoning workflows. Despite these substantial capability expansions, no benchmark data is available for either the current or previous window, making it impossible to assess whether performance on standard evaluation metrics has improved, degraded, or remained stable. The version number jump from 3.5 to 5.5 suggests a major release, and the breadth of new features indicates this is a significant platform evolution. Users gain access to substantially more functionality, particularly for applications requiring visual understanding, structured data extraction, or complex reasoning tasks. However, without benchmark comparisons, users cannot determine if core language understanding and generation quality has changed alongside these new capabilities.

Quality

—

Latency p50

—

Test runs

0

✓ Seven new capabilities added✓ Multimodal support introduced✓ Structured output modes available
Last automated test
Nov 22, 2026 · 14:02 UTC · Speed benchmark
P50 latency
7721 ms
P95 latency
7790 ms
Errors
0 / 6 runs
Last reviewed by Tokonomix Team·November 22, 2026