Claude Sonnet 4.5
claude-sonnet-4-5claude-sonnet-4.5sonnet-4-5sonnet-4.5- 🏢 Anthropic's balanced Sonnet model, released 2025, for coding and agents
- 🔧 Optimized for autonomous, long-horizon agentic and tool-use workflows
- 🧠 Substantial gains in reasoning and math over prior models
- 📏 200K-token context window for long documents and codebases
- 👁️ Multimodal with vision, function calling, and web search built in
- ⚡ Extended thinking adds enhanced reasoning for complex tasks
- 🎯 Vendor reports code-edit error rate fell from 9% to 0% versus Sonnet 4
- 🔒 Operates under Anthropic's AI Safety Level 3 protections
Anthropic PBC is an American artificial intelligence company headquartered in San Francisco. Structured as a public benefit corporation, the lab develops large language models under the Claude name, with a research emphasis on building reliable, steerable, and safety-focused AI…
Explore 16 more models by Anthropic →Claude Sonnet 4.5 is the "balanced" tier of Anthropic's Claude lineup, positioned between speed-and-cost efficiency and high capability. Anthropic describes it as optimized for software coding, agentic tasks where the model runs in a loop and uses tools, and computer use, with strong showings in reasoning and mathematics. It supports a 200K-token context window, and on this catalog it carries vision, code-optimization, function-calling, web-search, and extended-thinking reasoning capabilities.
Compared with its same-family predecessor Claude Sonnet 4, Anthropic reports meaningful generational improvement: its system card notes substantial capability gains in reasoning and mathematics, plus more comprehensive safety refusals. On the company's own internal code-editing benchmark, it states the edit error rate dropped from 9% on Sonnet 4 to 0%, alongside better tool success at lower cost. Anthropic also positions it for long-horizon agentic and tool-use workflows.
Within the family, Sonnet 4.5 was succeeded by Claude Sonnet 4.6, which Anthropic positions as a full upgrade across coding, computer use, and long-context reasoning, with a 1M-token context beta and pricing matching Sonnet 4.5. For the deepest reasoning and high-autonomy work, Anthropic points users toward its Opus tier, now led by Claude Opus 4.8.
Practically, Anthropic frames Sonnet 4.5 for software engineering, agents, cybersecurity, and financial analysis where sustained reasoning and reliable tool use matter, and it remains accessible across Anthropic's API and major cloud platforms.
This About section is AI-generated from public sources via VeniceStats + Venice inference, with no human editing. It may contain inaccuracies.
| Seller | Reputation↓ | Routing | Input $/M | Cached $/M | Output $/M | Categories | API |
|---|---|---|---|---|---|---|---|
| ▲ Apex Ant 0x73b4…e736 | 71.90 | #1 | $1.67 | $0.17 | $8.35 | chat,reasoning,coding,vision,multimodal,long-context,agents,tasks,legal,study,translate | openai-chat-completions |
| Venice.ai Proxy 0x1f22…18c9 | 70.55 | #3 | $1.875 | $0.1875 | $9.375 | chat,reasoning,coding,vision,multimodal,web-search | openai-chat-completions |
| Super Seeder 0xd19f…41f3 | 63.77 | #2 | $1.8563 | $0.1856 | $9.2813 | chat,coding,reasoning,vision,multimodal,tools | openai-chat-completions |
| Open Ant 0xe4f6…5bc4 | 50.00 | gated | $3.75 | $0.375 | $18.75 | chat,coding,reasoning,vision,multimodal,tools | openai-chat-completions |
| Fire Ant 🔥🐜 0xbe05…bc5d | 50.00 | gated | $3.0829 | $0.3083 | $15.4147 | — | — |
| NovaRoute AI 0xc50d…ed7b | 50.00 | gated | $0.5498 | $0.5498 | $2.7492 | chat,coding,code,reasoning,tasks,claude,sonnet,value,surplus,openai-compatible,low-cost,verified,github,response-auth,base-usdc,monitored | openai-chat-completions |
| Meridian AI 0x8c8c…06f5 | 44.01 | gated | $0.3645 | $0.3645 | $1.8225 | chat,coding,reasoning | openai-chat-completions |
| Relay Works 0x3040…7139 | 38.53 | gated | $1.40 | $0.14 | $7.00 | text-generation | openai-chat-completions |
| antseed-neon-puma-944e 0x6650…944e | 35.09 | gated | $1.2788 | $0.375 | $6.3938 | chat,coding,math | openai-chat-completions |
| Leftermute 0x388b…5389 | 28.39 | gated | $0.3409 | $0.3409 | $1.7044 | chat,coding,json,reasoning,tools | openai-chat-completions |
| Apex TEE Test 0xe672…7955 | 15.22 | gated | $0.5554 | $0.5554 | $2.777 | chat,reasoning,coding,vision,multimodal,long-context,agents,tasks,legal,study,translate | openai-chat-completions |
"Best price" and the seller table are live AntSeed catalog data (advertised $/1M tokens — or $ per generated image for unit-billed image models — not settled amounts). Reputation = buyer trust score (0-100, the AntSeed SDK's own formula). "Routing" = the SDK's default buyer routing (what the VPR desktop app ships with): a trust ≥ 60 gate on the effective reputation, then cheapest-first among routable sellers; live failover state (per-peer cooldowns) is buyer-side runtime and not included. Model knowledge (TLDR, provider, About) via the VeniceStats enrichment layer. Advertised catalog, not the model used in any specific purchase. "Usage on AntSeed" counts only settlements whose buyers share the per-model split on-chain (metadata v2/v3, opt-in), so every usage figure is a lower bound.