Noometry

Anthropic, proprietary

Claude Sonnet 4.6

Claude Sonnet 4.6 by Anthropic ranks 50th of 354 ranked models on the Noometry Index as of October 2026, with a score of 50.3. Its strongest category is writing & preference, where it ranks 22nd. API pricing starts at $3 per million input tokens and $15 per million output tokens, with a 1M-token context window.

Last verified

Noometry rank
#50 of 354
Index score
50.3
Evidence
Confirmed 57 results
Provider
Anthropic
Released
February 17, 2026
Weights
Proprietary
Reasoning
Yes
Context window
1M
Max output
128K
Input price
$3 / M
Output price
$15 / M
Blended price
$6 / M
Output speed
Not measured
Value
#191 of 219
Knowledge cutoff
August 2025
Input
text, image, pdf

Category scores

Each category score combines every public result we have in that category.

  1. Coding 46.3
  2. Agentic & Tool Use 39.1
  3. Reasoning 46.1
  4. Math 52.9
  5. Knowledge 51.7
  6. Multimodal 38.0
  7. Multilingual 54.4
  8. Instruction Following 77.4
  9. Long Context 45.3
  10. Writing & Preference 70.2
CategoryScoreRankResults
Coding46.3#677
Agentic & Tool Use39.1#288
Reasoning46.1#4510
Math52.9#493
Knowledge51.7#654
Multimodal38.0#682
Multilingual54.4#411
Instruction Following77.4#251
Long Context45.3#441
Writing & Preference70.2#225

Strengths and weaknesses

Categories where Claude Sonnet 4.6 places highest and lowest among the models ranked in each, with its score against that category's median.

Strongest categories

CategoryScorevs medianRank
Writing & Preference70.2+16.4#22 of 312, top 8%
Instruction Following77.4+6.1#25 of 305, top 9%
Reasoning46.1+22.5#45 of 350, top 13%

Weakest categories

CategoryScorevs medianRank
Multimodal38.0−0.5#68 of 128, top 54%
Knowledge51.7+14.4#65 of 314, top 21%
Coding46.3+7.6#67 of 340, top 20%

Closest competitors

The models ranked just above and below Claude Sonnet 4.6. When scores are this close, price and speed are often the better way to choose.

ModelRankScoreBlended $/MSpeed
Muse Spark#4650.6——Compare
Claude Opus 4.5#4750.5$1013Compare
Muse Spark 1.2#4850.3$2—Compare
MiMo-V2.6-Pro#4950.3$0.54—Compare
Muse Spark 1.1#5149.9$2—Compare
Claude Haiku 5.5#5249.5$0.20—Compare
GPT-5.1#5349.0$3.44—Compare
Grok 4.20 (Non-Reasoning)#5448.6$1.5661Compare

Sponsored placements are available on pages like this one. Advertise on Noometry

Benchmark results

Every published result we track, with its source. Bold rows are the ones used for ranking; where several exist we prefer independent runs over self-reported numbers.

Coding

BenchmarkScorePositionSettingSourceDate
SWE-bench Verified75.2%#14 of 32, top 44%Epoch AI2026-02-21
DeepSWE29.9%#28 of 29, top 97%highEpoch AI
FrontierCode24.3%#31 of 37, top 84%Epoch AI
LMArena WebDev1522#40 of 113, top 36%LMArena2026-10-08
SciCode46.8%#56 of 121, top 47%maxEpoch AI
WeirdML66.1%#23 of 119, top 20%mediumEpoch AI
LMArena Coding1504#17 of 294, top 6%LMArena2026-10-08
ALE-Bench1,327#21 of 105, top 20%mediumEpoch AI

Agentic & Tool Use

BenchmarkScorePositionSettingSourceDate
Terminal-Bench53.4%#15 of 41, top 37%Epoch AI
APEX-Agents43%#34 of 49, top 70%highEpoch AI
OSWorld 2.08.3%maxEpoch AI
OSWorld 2.09.3%#6 of 9, top 67%mediumEpoch AI
DeepResearch Bench54.9%#2 of 24, top 9%highEpoch AI
DeepResearch Bench50.4%lowEpoch AI
OSWorld72.1%Best of 8Epoch AI
ExploitBench23.6%#5 of 9, top 56%Epoch AI
GBAEval48.8%#8 of 23, top 35%Epoch AI
GDP.pdf18%#22 of 36, top 62%maxEpoch AI
LMArena Search1221#7 of 32, top 22%LMArena2026-08-24
Vending-Bench 27,204#15 of 60, top 25%Epoch AI

Reasoning

BenchmarkScorePositionSettingSourceDate
ARC-AGI-260.4%#27 of 83, top 33%highEpoch AI
ARC-AGI-258.3%maxEpoch AI
NYT Connections (extended)76.4%Lech Mazur benchmarks
NYT Connections (extended)48%Lech Mazur benchmarks
NYT Connections (extended)80.9%#33 of 91, top 37%high reasoningLech Mazur benchmarks
ARC-AGI-186.5%#34 of 83, top 41%highEpoch AI
ARC-AGI-186%maxEpoch AI
CritPt3.1%#64 of 134, top 48%maxEpoch AI
Chess Puzzles13%#72 of 129, top 56%32KEpoch AI2026-02-20
Chess Puzzles5%highEpoch AI2026-07-13
Chess Puzzles3%maxEpoch AI2026-08-06
Chess Puzzles8%mediumEpoch AI2026-07-13
Thematic Generalization76.3%#4 of 23, top 18%high reasoningLech Mazur benchmarks
LMArena Hard Prompts1484#25 of 297, top 9%LMArena2026-10-08
Mystery Game Puzzles14%Epoch AI2026-08-06
Mystery Game Puzzles16%#54 of 74, top 73%lowEpoch AI2026-08-06
DTBench89.9%#42 of 151, top 28%maxEpoch AI
LMCA46.5%#29 of 125, top 24%maxEpoch AI
Epoch Capabilities Index152.24#43 of 213, top 21%Epoch AI2026-02-17
ForecastBench59.6Epoch AI
ForecastBench62#3 of 72, top 5%16KEpoch AI

Math

BenchmarkScorePositionSettingSourceDate
OTIS Mock AIME 2024-202585.8%#69 of 173, top 40%32KEpoch AI2026-02-20
OTIS Mock AIME 2024-202575.6%highEpoch AI2026-07-13
OTIS Mock AIME 2024-202571.1%maxEpoch AI2026-08-06
OTIS Mock AIME 2024-202582.2%mediumEpoch AI2026-07-13
ProofBench45%#33 of 77, top 43%maxEpoch AI
LMArena Math1462#51 of 285, top 18%LMArena2026-10-08
FrontierMath (Feb 2025 set)32.4%#17 of 68, top 25%16KEpoch AI2026-02-20
FrontierMath Tier 4 (v1)8.3%#22 of 55, top 40%16KEpoch AI2026-02-21

Knowledge

BenchmarkScorePositionSettingSourceDate
GPQA Diamond87.4%#54 of 186, top 30%32KEpoch AI2026-02-20
GPQA Diamond83.3%highEpoch AI2026-07-13
GPQA Diamond78.8%maxEpoch AI2026-08-06
GPQA Diamond83.3%mediumEpoch AI2026-07-13
SimpleQA Verified35.5%#47 of 77, top 62%highEpoch AI2026-08-10
SimpleQA Verified32.8%maxEpoch AI2026-08-27
Vectara Hallucination Rate (lower is better)10.6%#61 of 96, top 64%Vectara Hallucination Leaderboard
LMArena Expert1500#26 of 273, top 10%LMArena2026-10-08

Multimodal

BenchmarkScorePositionSettingSourceDate
LMArena Vision1283#27 of 122, top 23%LMArena2026-10-09
Blueprint-Bench 26.7%#25 of 31, top 81%Epoch AI
LMArena Document1482#8 of 38, top 22%LMArena2026-09-13

Multilingual

BenchmarkScorePositionSettingSourceDate
LMArena Non-English1440#42 of 297, top 15%LMArena2026-10-08
LMArena Chinese1491#54 of 285, top 19%LMArena2026-10-08
LMArena French1465#44 of 223, top 20%LMArena2026-10-08
LMArena German1428#68 of 231, top 30%LMArena2026-10-08
LMArena Japanese1420#40 of 211, top 19%LMArena2026-10-08
LMArena Korean1411#41 of 213, top 20%LMArena2026-10-08
LMArena Russian1440#53 of 283, top 19%LMArena2026-10-08
LMArena Spanish1464#30 of 226, top 14%LMArena2026-10-08

Instruction Following

BenchmarkScorePositionSettingSourceDate
LMArena Instruction Following1475#22 of 298, top 8%LMArena2026-10-08

Long Context

BenchmarkScorePositionSettingSourceDate
LMArena Longer Query1479#26 of 291, top 9%LMArena2026-10-08

Writing & Preference

BenchmarkScorePositionSettingSourceDate
LMArena Text1458#35 of 297, top 12%LMArena2026-10-08
LMArena Creative Writing1435#45 of 295, top 16%LMArena2026-10-08
EQ-Bench Creative Writing1810#22 of 115, top 20%EQ-Bench
EQ-Bench 41207#16 of 28, top 58%EQ-Bench
LMArena Multi-Turn1464#36 of 295, top 13%LMArena2026-10-08

API pricing by provider

RouteInput $/MOutput $/MCached input $/MChecked
anthropic$3$15$0.302026-10-10
azure$3$15$0.302026-10-10
bedrock$3.30$16.50$0.332026-10-10
openrouter$3$15$0.302026-10-10
vertex$3$15$0.302026-10-10

All Anthropic API prices → Estimate your cost →

Compare Claude Sonnet 4.6

Other Anthropic models

Frequently asked questions

How good is Claude Sonnet 4.6?

Claude Sonnet 4.6 by Anthropic ranks 50th of 354 ranked models on the Noometry Index as of October 2026, with a score of 50.3. Its strongest category is writing & preference, where it ranks 22nd. API pricing starts at $3 per million input tokens and $15 per million output tokens, with a 1M-token context window.

How much does Claude Sonnet 4.6 cost?

Claude Sonnet 4.6 costs $3 per million input tokens and $15 per million output tokens on Anthropic's own API, with cached input at $0.30.

What is Claude Sonnet 4.6's context window?

Claude Sonnet 4.6 accepts up to 1M tokens of input and can write up to 128K tokens in one response.

Is Claude Sonnet 4.6 open source?

No. Claude Sonnet 4.6 is proprietary and available only through Anthropic's API and partner platforms.

What are Claude Sonnet 4.6's strengths and weaknesses?

Relative to other ranked models, Claude Sonnet 4.6 places best in writing & preference, instruction following, reasoning and lowest in multimodal, knowledge, coding.

What is Claude Sonnet 4.6 best at?

Its best category is writing & preference, where it ranks 22nd on Noometry.