Compare Start here

AnthropicLanguage model

Claude Sonnet 4.6

Anthropic language model.

Compare head-to-head Race it in 3D

Overall Intelligence

#41

of 61
49.3 pts behind #1

0

races led

10

weeks at #1

41.8

score

Released: Feb 17, 2026; 8 mo ago
Context: 1M; 128K max output
Price: $3 / $15; per 1M in/out · $6.00 blended · OpenRouter
Speed: —; not measured
Modalities: text · image; outputs text
Weights: Closed; API and apps

The 30-second version

Anthropic language model.

Key takeaways

1Anthropic language model.

Where it ranks

Latest real results · click any race to watch it

Sonnet 4.6Best on each axis

Overall

41Overall+49.3 pts

Coding

36Coding+59.5 pts 40Frontend & UI+284 Elo 15Coding Q&A+43 Elo 41React Web Apps+343 Elo 11Debugging & Repair+8.3 pts 40Terminal & DevOps+60.6 pts

Hard Reasoning

36Hard Reasoning+59.3 pts 49Frontier Questions+27.8 pts 27ARC-AGI-2+34.6 pts

Math

42Math+62.1 pts 49Competition Math+17.8 pts 37Math Chat+74 Elo

Science

48Science+62.6 pts 45Graduate Science QA+13.1 pts 37Medicine & Health+103 Elo 42Scientific Coding+16.8 pts 28Science Q&A+73 Elo

Writing & Language

31Writing & Language+43.0 pts 35Creative Writing+98 Elo 35Professional Writing+91 Elo 32Multilingual+82 Elo 21Instruction Following+60 Elo 29Conversation+92 Elo

Multimodal

16Multimodal+34.8 pts 24Image Understanding+42 Elo 8Documents & Charts+34 Elo

Tool Use & Agents

17Tool Use & Agents+54.1 pts 24Tool Use+16.9 pts 6Computer Use+22.1 pts 11Web Search+41 Elo

Long Context

36Long Context+49.3 pts 36Long-Document Reasoning+8.7 pts

Speed & Cost

55Intelligence per Dollar+615.1 pts

Trust & Factuality

46Trust & Factuality+85.1 pts 41Factuality+40.5 pts

Overall Intelligence over time

#1 frontierSonnet 4.6Grok 4.7MiMo 2.6 Flash

Update history

The distilled version first; open any release for the deeper breakdown.

The Claude Sonnet family

3.7 SonnetFebruary 2025 Sonnet 4.5September 2025 Sonnet 4.6February 2026 Sonnet 5June 2026 Sonnet 5.5September 2026

60-second briefing

Sonnet 4.6, explained

Animated summary · captions from the auto-written script

From Anthropic's own guide, distilled

Prompt engineering and Claude best practices

Be explicit about the outcome you want; newer Claude models follow instructions literally rather than guessing at hidden intent.
Explain why a rule exists. Context about the goal lets the model generalise the rule to cases you did not list.
Wrap documents, examples and data in XML-style tags so instructions and material never blur together.
Put long documents at the top of the prompt and your question at the end; recall and grounding both improve.
For agentic tasks, state when to act versus when to ask, and ask for a short plan before large edits.

All official guides, distilled →

About Anthropic

Safety-focused lab building the Claude family of models.

Optimising for

Reliable long-horizon agentic work (especially coding and computer use), honest and calibrated answers, and strong writing, with safety research shipped alongside capability.

San Francisco, US · founded 2021

Closest rivals

Sonnet 4.6 is 41st in overall intelligence; leader is Claude Opus 5.5.