AI Model Release Tracker

Claude Sonnet 5

Released

Claude Sonnet 5 is an AI model released by Anthropic on Tuesday, Jun 30 2026, 21 days after Claude Mythos 5. It has a 1M token context windowHow much information the model can work with in one request, measured in tokens — pieces of words or other data. This includes your question, attachments, conversation and reply. Limits can vary by model variant and provider.. Benchmark results (shown below) cover BullshitBench v2, Gray Swan IPI, SWE-Bench Pro, SWE-Bench Verified, SWE-Bench Multilingual, SWE-Bench Multimodal, and 21 more.

API pricing

InputWhat you pay for everything you send the model — your question, plus any documents or earlier conversation you include with it.
$2.00
Cached inputA reduced rate for text you send over and over. If every request starts with the same instructions or the same document, the provider keeps a copy ready and charges less to read it again.
$0.20
5 min cache writeA one-off charge for storing text so later requests can reuse it at the cheaper cached rate. This option keeps it for five minutes.
$2.50
1 hr cache writeA one-off charge for storing text so later requests can reuse it at the cheaper cached rate. This option keeps it for an hour, so it costs more than the five-minute one.
$4.00
OutputWhat you pay for the text the model writes back. It is normally the dearer half: producing an answer costs more than reading one.
$10.00
  • Adaptive thinking is on by default.
USD per 1M tokensEvery price here is for one million tokens. A token is roughly three-quarters of a word, so a million tokens is about 750,000 words of text.Verified October 19, 2026Read from platform.claude.com. Click the check to open the source.

Available from

Amazon Bedrockglobal$2.00$10.001M
Anthropic—$2.00$10.001M
Azureglobal$2.00$10.001M
Claude Platform on AWS—$2.00$10.001M
Googleglobal$2.00$10.001M
Amazon Bedrockeu-west-1$2.20$11.001M
Amazon Bedrockus-east-1$2.20$11.001M
Azureus$2.20$11.001M
Googleeu$2.20$11.001M
Googleus$2.20$11.001M
USD per 1M tokensEvery price here is for one million tokens. A token is roughly three-quarters of a word, so a million tokens is about 750,000 words of text.Checked November 12, 2026Third-party rates from openrouter.ai, not the lab’s own list price.

Benchmarks

Coding

SWE-Bench ProAgentic coding — Can the AI fix real bugs in real software? It's handed actual problems from open-source projects and has to write code that genuinely solves them. Higher is better.
63.2%
#15 of 42
SWE-Bench VerifiedCoding — Real coding tasks pulled from open-source projects — the AI has to find and fix actual bugs. A human-checked version of the original SWE-Bench. Higher is better.via BenchLM
85.2%
#7 of 80
SWE-Bench MultilingualMultilingual coding — Like SWE-Bench, but the coding problems span many programming languages, not just one. Tests how broadly the AI can code. Higher is better.
78.3%
#12 of 26
SWE-Bench MultimodalMultimodal coding — Real bug reports that arrive with pictures attached — a screenshot, a mockup, a page rendering wrongly — so the AI has to read the image as well as the code to work out what to fix. Higher is better.
28.1%
#10 of 12
CursorBench 4.0Agentic coding — Cursor's own test of coding agents on ambiguous, multi-file tasks taken from real Cursor sessions — editing, refactoring, investigating a codebase, understanding what the user meant, managing jobs and following a design. Cursor runs each model at several reasoning efforts; each release here carries the score of its best listed effort. Scores aren't comparable with earlier CursorBench versions. Higher is better.via CursorBench
34.1%
#17 of 18
FrontierCode v1.1 (Main)Agentic coding — A set of very hard, frontier-difficulty coding tasks an AI agent has to complete end to end. The score is the share of tasks in the main split it solves. Higher is better.
42.4%main split
#10 of 11
Next.js EvalsNext.js coding — Vercel's open eval of how well AI coding agents build and migrate real Next.js apps — measured as the share of tasks the agent completes successfully. Higher is better.via Next.js Evals
82%
#13 of 36
Supabase EvalsSupabase coding — Supabase's own open benchmark: a coding agent is dropped into a real Supabase project and asked to do real work — set up a schema, fix a broken security policy, debug an Edge Function — and every run is checked against a live Supabase stack. This is the headline number, where the agent has Supabase's own skills loaded, as most people building on Supabase would. The score is the share of scenarios it got right. Higher is better.via Supabase Evals
79.7%with skills
75.4%no skills
with skills
#7 of 11
no skills
#7 of 11

Terminal & CLI

Frontier-Bench v0.1Agentic computer work — A hard, ever-evolving set of real computer tasks — coding, system administration, data work, and more — that an AI agent has to complete on its own. Run by the Harbor / Laude Institute team as the successor to Terminal-Bench (v0.1 is the first release of the task set). The score is the share of tasks solved. Higher is better.
14.6%
#7 of 9
Terminal-Bench 4.0Agentic terminal coding — Can the AI work in a command-line terminal — running commands and finishing technical setup tasks the way a developer would? Version 4.0 recalibrated how much time, CPU and memory each task gets, removed eight tasks and fixed nineteen, so fewer runs fail for reasons that have nothing to do with the model. Scores are not comparable with earlier versions. Higher is better.via Terminal-Bench
12.42%
#33 of 36
Terminal-Bench 2.1Agentic terminal coding — Can the AI work in a command-line terminal — running commands and finishing technical setup tasks the way a developer would? Higher is better.
80.4%
#20 of 40

Agentic & tool use

BrowseCompWeb browsing — Can the AI browse the web and track down hard-to-find answers? Higher is better.via BenchLM
84.7%
#12 of 35
OSWorld 2.1Agentic computer use — An agent operates a desktop across 108 tasks. Each completed checkpoint earns part of the task’s credit, so a model can score even when it only finishes part of a task.
57%full set
#5 of 5
OSWorld-VerifiedAgentic computer use — Can the AI actually operate a computer — clicking, typing, and using real apps — to finish tasks on its own? Higher is better.
81.2%
#7 of 32
AutomationBenchBusiness workflows — Tests whether the AI can run real multi-step business workflows — the kind of end-to-end office processes companies want to automate — from start to finish. Higher is better.
13.5%
#30 of 32

Reasoning & science

Humanity's Last ExamMultidisciplinary reasoning — Humanity's Last Exam — extremely hard expert questions across many subjects, written so you can't just look up the answer. “No tools” means the AI answers on its own. Higher is better.
43.2%no tools
57.4%with tools
no tools
#14 of 48
with tools
#16 of 60
BioMysteryBenchBiology — Real unsolved-style biology puzzles — the AI has to reason its way to an answer the way a research biologist would. The “hard” split contains the toughest cases. Higher is better.
32.9%hard
87.5%human solved
hard
#10 of 10
human solved
#6 of 10

Knowledge work

GDPval-AA v2.1Knowledge work — economically valuable knowledge work (v2.1, Crowd-BT Elo fit)
1449
#8 of 9
GDPval-AA v2Knowledge work — economically valuable knowledge work (v2, re-based Elo)
1618
#10 of 25
AA-Briefcase v1.1Knowledge work — Artificial Analysis agentic office-work eval (Elo, v1.1 rating fit)
1359
#12 of 13

Healthcare

HealthBench ProfessionalHealth — physician-graded conversations, scored before adjustment for answer length
62.4%raw
57.8%length-adjusted
raw
#8 of 9
length-adjusted
#8 of 10

Multimodal

ChartographyChart tasks — The same chart-centred test with no tools: the AI has to read each chart unaided. Scores run far lower than the with-tools version, so read the two as separate tests. Higher is better.
15.6%no tools
#14 of 15

Robustness

BullshitBench v2Nonsense detection — Given a confidently-worded but nonsensical prompt, does the AI spot that it makes no sense and push back — instead of playing along and inventing an answer? The score is how often it clearly called out the nonsense. Higher is better.via BullshitBench
80%
#7 of 100
Gray Swan IPIPrompt injection robustness — Attackers hide malicious instructions inside content the AI reads — a web page, an email, a document — and try to hijack what it does. Gray Swan's indirect prompt injection benchmark measures how often such an attack succeeds when the attacker gets a single try. Lower is better.
0.6%k = 1
4.7%k = 10
5.9%k = 15
k = 1
#4 of 13
k = 10
#4 of 13
k = 15
#6 of 15

Community preference

threejsevalCommunity preference (Three.js) — Every model gets the same prompt — "the Eiffel Tower", "a glass fishbowl", "a robot arm picking toys into a box" — and builds a 3D scene in Three.js. Real people then see two scenes side by side, names hidden, and vote for the one they prefer. The votes become a chess-style Elo rating on threejseval.com, averaged across all the prompts. It measures whether the scene looks and moves right to a human eye, not whether the code passes a test. Higher is better.via threejseval
1215
#25 of 29

Source: CursorBench, retrieved 28 September 2026 · Source: Terminal-Bench, retrieved 30 August 2026 · Source: BenchLM, retrieved 24 August 2026. Other sources are identified on the linked benchmark pages.

Compare Claude Sonnet 5 with

Claude Sonnet 5

Suggested comparisons

Claude Sonnet 5vsClaude Fable 5.2Claude Sonnet 5vsGPT-6.1 TerraClaude Sonnet 5vsGemini 4 HydrogenClaude Sonnet 5vsMuse EmberClaude Sonnet 5vsGrok 4.7Claude Sonnet 5vsDeepSeek-V4.1-Flash-1017Claude Sonnet 5vsMistral Large 4Claude Sonnet 5vsKimi K3.1 ProClaude Sonnet 5vsGLM-5.3-FlashClaude Sonnet 5vsQwen4-Max PreviewClaude Sonnet 5vsNemotron 3.5 Bolt E4B 4GB

About

Claude Sonnet 5 launched on June 30, 2026 as Anthropic's next Sonnet model for planning, tool use, coding and knowledge work. The company positioned it as approaching Opus 4.8's capability on many agent tasks at a lower price. It became the default for Free and Pro users and was available across paid Claude plans and the API.

The release emphasized completing work through browsers and terminals, with effort settings providing a way to vary computation by task. Anthropic reported improvements over Sonnet 4.6 in agentic performance and safety evaluations, while noting lower cybersecurity capability than its contemporary Opus models. Its role was a broadly accessible model for multi-step tasks, rather than a restricted frontier-security variant.

Frequently asked questions

Claude Sonnet 5 was released by Anthropic on Tuesday, Jun 30 2026.

All Anthropic releases

33 tracked