Skip to content
AI Primer

Claude Sonnet 4.6 is a specific Anthropic Claude Sonnet model release announced on Feb 17, 2026, described as a full upgrade across coding, computer use, long-context reasoning, agent planning, knowledge work, and design, with a 1M token context window in beta.

Pricing

Official site · Aug 10, 2026, 7:24 AM
Input / 1M
$3.00
Output / 1M
$15.00
Cached input / 1M
$0.30

All prices are USD per million tokens (MTok). For Claude Sonnet 4.6, 5-minute cache writes are $3.75/MTok, 1-hour cache writes are $6/MTok, and cache hits & refreshes are $0.30/MTok.

Anthropic's official Claude Platform pricing documentation lists Claude Sonnet 4.6 model pricing in USD per million tokens: $3/MTok base input, $15/MTok output, and $0.30/MTok cache hits & refreshes. It also lists prompt cache writes at $3.75/MTok for 5 minutes and $6/MTok for 1 hour.

View source

Model Intelligence

Context window
1,000,000 tokens
Arena ranking
24
Benchmarkable
Yes
Model level
release
Intelligence Index
23.9
Math Index
21
MMLU Pro
0.8
GPQA
0.66
HLE
0.04
LiveCodeBench
0.39
SciCode
0.38
MATH-500
0.85
AIME
0.22
AIME 2025
0.21
IFBench
0.44
LCR
0.5
TerminalBench Hard
0.21
TAU2
0.5

Recent stories

6 linked stories
releaseSECONDARY2026-07-06
Tencent releases Hy3, a 295B MoE model under Apache license

Tencent released Hy3 with 21B active parameters, a 256K context window, BF16/FP8 weights, and day-one vLLM/SGLang support. Kilo Code, Nous Portal, and OpenRouter also made it free for limited windows.

newsSECONDARY2026-04-10
GLM-5.1 ranks #3 on Code Arena

Arena ranked GLM-5.1 third on Code Arena and first among open models, putting it on par with Claude Sonnet 4.6 and within about 20 points of the overall lead. The update gives the open model a new frontier coding benchmark after its initial release and hosting wave.

releaseSECONDARY2026-04-09
Anthropic adds beta advisor tool to Messages API for Opus calls

Anthropic added a beta advisor tool to the Messages API so Sonnet or Haiku can call Opus mid-run inside one request. Anthropic says Sonnet plus Opus scored 2.7 points higher on SWE-bench Multilingual while cutting per-task cost 11.9%.

newsPRIMARY2026-03-23
LLM Debate Benchmark ranks Sonnet 4.6 first across 1,162 side-swapped debates

LLM Debate Benchmark ran 1,162 side-swapped debates across 21 models and ranked Sonnet 4.6 first, ahead of GPT-5.4 high. It adds a stronger adversarial eval pattern for judge or debate systems, but you should still inspect content-block rates and judge selection when reading the leaderboard.

newsSECONDARY2026-03-14
Claude Opus 4.6 ranks 78.3% on MRCR v2 at 1M tokens

Third-party MRCR v2 results put Claude Opus 4.6 at a 78.3% match ratio at 1M tokens, ahead of Sonnet 4.6, GPT-5.4, and Gemini 3.1 Pro. If you are testing long-context agents, measure retrieval quality and task completion, not just advertised context window size.

releasePRIMARY2026-03-13
Anthropic launches 1M-token context for Opus 4.6 and Sonnet 4.6 at flat pricing

Anthropic made 1M-token context generally available for Opus 4.6 and Sonnet 4.6, removed the long-context premium, and raised media limits to 600 images or PDF pages. Use it for retrieval-heavy and codebase-scale workflows that previously needed beta headers or special long-context pricing.

AI PrimerAI Primer

Your daily guide to AI tools, workflows, and creative inspiration.

© 2026 AI Primer. All rights reserved.