LLManthropic

Anthropic: Claude Sonnet 5

Sonnet 5 is Anthropic's most capable Sonnet-class model, with frontier performance across coding, agents, and professional work. It supports adaptive thinking with selectable reasoning effort levels (low, medium, high, max,...

Anyone in the Space can @-mention Anthropic: Claude Sonnet 5 with the team's shared context - pooled credits, one chat, one memory.

All models

Starter is free forever - 1 Space, 100 credits/month, 1 MCP. No card.

Verdict

Claude Sonnet 5 delivers Anthropic's strongest reasoning and coding performance in the Sonnet tier, with a 1M token context window that handles entire codebases or long documents in a single pass. It costs 5x more on output than Sonnet 4, making it expensive for high-volume generation, but the jump in accuracy on complex tasks justifies the premium when quality matters more than speed. Reach for this when you need Claude's safety guardrails plus reasoning depth that rivals Opus-class models at a fraction of the cost.

Best for

Multi-file codebase refactoring and analysis
Long-context document synthesis across 100+ pages
Complex reasoning tasks requiring chain-of-thought
Vision tasks with detailed image analysis
Workflows where safety and refusal behavior matter

Strengths

The 1M token window lets you load entire repositories or multi-document sets without chunking, and Sonnet 5 maintains coherence across that full span better than earlier Sonnet versions. Coding tasks see a measurable step up from Sonnet 4—fewer hallucinated function names, better adherence to project conventions. Vision capabilities handle screenshots and diagrams with more precise detail extraction than previous Sonnet releases. Anthropic's constitutional AI training means fewer false refusals on legitimate business use cases compared to Sonnet 3.5.

Trade-offs

Output pricing at $10/Mtok makes this 5x more expensive than Sonnet 4 for generation-heavy workflows like drafting long reports or bulk content creation. Latency sits higher than GPT-4o or Gemini Flash for equivalent tasks, so real-time chat or low-latency APIs will feel slower. While reasoning improved, it still trails Claude Opus 4 on the hardest logic puzzles and multi-step math, so don't expect Opus-tier performance at Sonnet pricing. No function calling or structured output modes yet, limiting integration patterns that competitors support natively.

Specifications

Provider: anthropic
Category: llm
Context length: 1,000,000 tokens
Max output: 128,000 tokens
Modalities: text, image, file
License: proprietary
Released: 2026-06-30

Pricing

Input: $2.00/Mtok
Output: $10.00/Mtok
Model ID: anthropic/claude-sonnet-5

Per-token prices show what the model costs upstream. On Switchy your team draws from one shared org credit pool - one plan, one balance for everyone.

Team cost calculator

Seats5 peopleMessages / seat / day80Avg turn size2 ktokOutput share30 %

Estimated monthly spend

$77.44

17.6M tokens / month
5 seats · 80 msgs/day

Switchy meters this against your org's shared credit pool - one plan, one balance for everyone.

Providers

Provider	Context	Input	Output	P50 latency	Throughput	30d uptime
anthropic	1000k	$2.00/Mtok	$10.00/Mtok	—	—	—

Performance

Performance snapshots are collected daily. Check back after the next ingestion run.

Benchmarks

Public benchmark scores are not available yet for this model. Check back after the next ingestion run.

Works well with

Top MCPs

Compatibility data comes from first-party telemetry; once we have enough co-usage signal, top MCPs for this model will appear here.

How Switchy teams use it

Not enough Spaces have used this model yet to share anonymised team stats. We wait for at least 50 distinct Spaces per week before publishing any aggregate.

Starter prompts

Refactor Legacy Codebase

I'm pasting a full codebase below (15 files, ~8000 lines). Identify the three highest-impact refactorings that would improve maintainability without requiring a rewrite. For each, explain the current problem, the proposed change, and which files are affected.

Open in a Space →

Synthesize Research Papers

I've attached five research papers on [topic]. Create a synthesis table showing: (1) each paper's main claim, (2) methodology differences, (3) where findings agree or conflict. Then write a two-paragraph summary of the current consensus.

Open in a Space →

Debug Complex Logic Error

This function should return X but returns Y when input is Z. I'm pasting the full call stack and related functions. Walk through the execution step-by-step, showing variable states at each level, and identify where the logic breaks.

Open in a Space →

Extract Data from Screenshots

I'm attaching three screenshots of an invoice. Extract all line items into a JSON array with fields: description, quantity, unit_price, total. Flag any amounts that look incorrect or don't sum properly.

Open in a Space →

Draft Technical Specification

Based on these product requirements [paste requirements], write a technical specification covering: system architecture, API endpoints with request/response schemas, database schema, and error handling strategy. Use consistent naming conventions throughout.

Open in a Space →