AnthropicGeneralReleased 28 Sept 2026

Claude Sonnet 5.5

Scoped coding & writing. Everyday coding and agents; a writing-workflow candidate.

54.5
· max
$0.30100K-in + 10K-out request
$2.00Input / 1M tokens
$10.00Output / 1M tokens
144 tok/sOutput speed
Watch for: At max effort its benchmark task cost (~$10) matches Opus 5.5, offsetting the token-price advantage.
Estimate monthly cost

Which reasoning effort setting?

controls how long Claude Sonnet 5.5 thinks before answering. Token prices never change; higher settings spend more and raise task cost.

Recommended start: Max
2030405060LowMediumHighXhighMax↑ INTELLIGENCE INDEXBenchmark cost per task →

Start at Max. The top setting is also the best value here. Rule: cheapest setting within 2 points of the model’s best score.

54.5Index score
$10.02Per task
19.6×vs Low
● Thinking 206K● Answer 79Ktokens / task

At Max, Sonnet 5.5 thinks for about 206K tokens per task, then writes about 79K more as its answer: $10.02 per task, 19.6× Low, for +28.9 points.

Output rate: $10.00/1M at every effort level. Thinking tokens are billed as output.

EffortIndexPer taskThinkingAll output× LowTime
Low25.6$0.518.6K20.9K1.00×2.6 min
Medium30.8$0.7814.4K30.7K1.53×3.7 min
High41.5$1.6431.7K58.4K3.20×7.1 min
Xhigh49.9$4.0979K122K8.01×14.4 min
MaxStart here54.5$10.02206K285K19.61×33 min

Per task of the Artificial Analysis Intelligence Index: the average over its nine evaluations, counting input, cache, and output tokens at list prices. These are long, agent-style benchmark tasks, much longer than a chat message.

Where Claude Sonnet 5.5 sits on the frontier

Intelligence Index against one 100K-input, 10K-output request.

Open full interactive map
203040506070$0.01$0.02$0.05$0.1$0.2$0.5$1$2↑ INTELLIGENCE INDEXCost of one request: 100K tokens in, 10K out (log scale) →Pareto frontierSonnet 5.5

Pricing, limits, speed & capabilities

Pricing

Standard rates per 1M tokens$2.00 in · $0.10 cached · $10.00 out
Long-prompt thresholdSame rate across the window
$1.00 in · $5.00 out (50% off)
None offered

Limits & modalities

1,000,000 tokens
Max output128,000 tokens
Inputs → outputText + image → text
AccessDirect API; app access and subscription limits vary.

Speed & capabilities

Output speed144 tokens/s
Time to first answer (incl. thinking)265 s
Tool callingYes
Structured outputYes
Computer useYes