Claude Opus 5 & Sonnet 5 Now Available: Near-Frontier Intelligence at Every Tier
Anthropic's Claude Opus 5 and Claude Sonnet 5 are now live on Multos. Opus 5 delivers near-Fable-5 intelligence at half the price with 79.2% on SWE-bench Pro, while Sonnet 5 brings Opus-class agentic capabilities to Starter-tier users at $3/$15 per million tokens.
Two major Anthropic models are now live on Multos. Claude Opus 5 — released today, July 24, 2026 — is Anthropic's strongest Opus-class model ever, closing the gap to their flagship Fable 5 at half the price. Claude Sonnet 5, which launched June 30, is the most agentic Sonnet to date and brings near-Opus quality to a far lower price point. Both are available immediately across Multos web, CLI, and API.
Claude Opus 5: The Highlights
Opus 5 combines a 1-million-token context window, 128K synchronous output (300K via batch), a five-level effort control (low/medium/high/xhigh/max), and the strongest benchmark profile of any Opus model. It scores 79.2% on SWE-bench Pro (up 10 points from Opus 4.8), 44.4 on Frontier-Bench at xhigh effort, 90.8% on BrowseComp web research, and 1,861 Elo on GDPval-AA v2 professional knowledge work — beating both GPT-5.6 Sol and Fable 5 on that last metric. Anthropic describes it as coming close to Fable 5's frontier intelligence at half the price.
Claude Sonnet 5: Opus-Class Agentic Power at Sonnet Pricing
Sonnet 5 ships with a 1M-token context window, always-on adaptive thinking, and a Terminal-Bench 2.1 score of 80.4% (up 13 points from Sonnet 4.6). It scores 63.2% on SWE-bench Pro, 57.4% on Humanity's Last Exam with tools, and 1,618 on GDPval-AA v2 — actually edging out Opus 4.8 on professional knowledge work. The key behavioural change: Sonnet 5 finishes multi-step agent tasks reliably rather than stalling or asking for permission mid-chain.
Best For
Claude Opus 5: Multi-repository refactoring, migration planning, debugging with test feedback, deep research with tool use, long document analysis, mathematical reasoning, and any task where sustained accuracy at scale matters. Highest-difficulty coding problems where the 6-point SWE-bench gap over Sonnet is worth the price premium. Claude Sonnet 5: Day-to-day agentic coding, terminal operations, tool-use workflows, professional knowledge tasks, and high-volume development work. The best value model in the Claude lineup — near-Opus quality at roughly one-third the cost for most workloads.
Specs at a Glance
Both models: 1M token context window, 128K max output, text + image input, text output, knowledge cutoff early 2026. Opus 5: $5/$25 per million tokens (input/output). Five effort levels. Available on Elite tier. Sonnet 5: $3/$15 per million tokens (input/output). Always-on adaptive thinking. Available on Pro tier and above.
How They Compare
Opus 5 vs GPT-5.6 Sol: Opus leads on SWE-bench Pro (79.2 vs 47.5) and ties on BrowseComp (90.8 vs 90.4). Cheaper output ($25 vs $30 per MTok) and no long-context surcharge. Sonnet 5 vs Opus 4.8: Sonnet 5 matches or beats Opus 4.8 on knowledge work (1,618 vs 1,615 GDPval) and Humanity's Last Exam with tools (57.4 vs 57.9) while costing 40% less. Sonnet 5 vs Sonnet 4.6: +13 points on Terminal-Bench, +5 on SWE-bench Pro, +10 on HLE with tools. A generational leap in agentic capability.
What We Replaced
With the addition of Sonnet 5 and Opus 5, we've retired Claude Haiku 4.5, Sonnet 4.6, Opus 4.5, and Opus 4.6 from the platform. Sonnet 5 replaces both Haiku 4.5 and Sonnet 4.6 as the go-to Anthropic model for most tasks — it's faster, smarter, and cheaper than running the old two-model split. Opus 5 replaces Opus 4.6 as the new flagship. Opus 4.7 and 4.8 remain available for users who prefer them.
Ready to Get Started?
Try Claude Opus 5 or Sonnet 5 now — select them from the model dropdown or let Auto Router assign the best model for your task.
Start Building FreeRelated Articles
Kimi K3 Now Available: The World's Largest Open-Weight Model on Multos
Moonshot AI's 2.8-trillion-parameter Kimi K3 is now live on Multos. A frontier-class open model with a 1M token context window, ranked #1 on Frontend Code Arena and #4 overall on Artificial Analysis Intelligence Index.
Read more →New Models: Gemini 3.6 Flash & Gemini 3.5 Flash-Lite Now Available
Two new Google Gemini models are now live on Multos — Gemini 3.6 Flash for Pro users and Gemini 3.5 Flash-Lite for Starter and above.
Read more →Auto Mode: How Smart Routing Stretches Your Token Budget
Auto mode picks the cheapest capable model per task. Manual selection gives you control with transparent cost multipliers.
Read more →