anthropic

Claude Opus 5.5

textimagefile
CAPABILITY & COST

What else is worth comparing?

Claude Opus 5.590.6% mean scoreGPQA Diamond · 2026-09-22
$9.001M input + 250k output tokens
Loading chart
This modelSelected alternativesOther dots: models with both score and price
HIGHER POINT SCORE · COST WITHIN +15%

Claude Sonnet 5.5

$4.50per workload95.6% mean score

+5.0 pp · $4.50 less
Evaluated / snapshot 2026-09-29

Compare both
NEARBY POINT SCORE

GPT-6 Luna

$0.225per workload90.5% mean score

-0.1 pp · $8.775 less
Evaluated / snapshot 2026-09-22

Compare both
Compare this selection

Same test and configuration; differences in point scores are not proof of a meaningful advantage. Nearby score does not mean equivalent capability.

INDEPENDENT EVALUATIONS

Results in context

Where does it sit?

Five nearest ECI estimates · current snapshot, not release-time scores.

Loading chart

Overlapping intervals do not establish a clear winner. ECI combines source-selected benchmarks; points are not percentages. Snapshot 2026-10-24.

Test by test

Loading chart
GPQA Diamond · max · 2026-09-22 · 21 included resultsFrontierMath tiers 1–3 · v2 · max · 2026-09-22 · 22 included resultsSimpleQA Verified · max · 2026-09-22 · 20 included resultsOTIS Mock AIME 2024–2025 · max · 2026-09-22 · 21 included results

Each row is a different test, not a common difficulty scale. “Max” does not mean equal compute across providers.

What changes your bill?

Illustrative token baskets, not measured tasks.

Loading chart
Short exchange1K input + 250 output tokens$0.009
Long answer4K input + 4K output tokens$0.096
Document analysis100K input + 2K output tokens$0.44
Repeated workload1M input + 250K output tokens$9.00

List prices, captured 2026-10-25. Excludes caching, batch discounts, tools and retries. Token baskets are not estimates of a specific task’s token use.

Within the same developer

Five lowest listed workload costs, plus this model.

Loading chart

Fixed basket: 1M input + 250k output. Logarithmic positions; labels show actual USD. Lower cost does not imply equivalent capability.

PUBLISHED LIMITS

Room to read
Room to respond

Context and maximum output are separate limits, not a quality score. They are not added together.

Context window1Mtokens
Maximum output128Ktokens

Published specifications

Context window1Mtokens
API input price$4USD / million tokens
API output price$20USD / million tokens
Maximum output128Kmaximum tokens per response
Output modalitiestextpublished output types
Supported parameters11API parameters listed

Model information

Provider
anthropic
Input modalities
text, image, file
Output modalities
text
Maximum output
128,000 tokens
Supported parameters
include_reasoning, max_completion_tokens, max_tokens, reasoning, reasoning_effort, response_format, stop, structured_outputs, tool_choice, tools, verbosity
Added to catalogue
2026-09-22
Knowledge cutoff
Not available
Catalogue snapshot
2026-10-25

Prices are per million tokens. The workload estimate uses 1,000 input and 250 output tokens. Context describes request capacity, not model quality.

Statvero