Claude Sonnet 3.5 launched on June 20, 2024 and instantly became one of the most talked-about models of its time: it beat the flagship Claude 3 Opus across almost every benchmark while running twice as fast and costing far less. For many people it was their first encounter with Claude. Since then, Anthropic has shipped several new Sonnet generations and retired 3.5 entirely — it's gone from the API and every catalog.
This article covers what earned Claude 3.5 Sonnet its reputation, whether you can still use it, and which Claude Sonnet version is current in 2026.
What made Claude 3.5 Sonnet great
At release, Claude 3.5 Sonnet set new industry records: graduate-level reasoning (GPQA), undergraduate-level knowledge (MMLU), and coding (HumanEval). In Anthropic's internal coding evaluation it solved 64% of tasks versus 38% for Claude 3 Opus — at twice the speed and a mid-tier price.
What it brought to the industry:
- A leap in coding. The model could write, edit, and run code on its own, and handled legacy migrations with confidence. It kicked off much of the AI coding assistant boom.
- Strong vision. The best vision model of its day: charts, graphs, and text transcription from imperfect photos.
- Artifacts. With 3.5 Sonnet came a window where code and documents could be edited right next to the conversation — an idea almost every chat service later adopted.
- The "smarter than the flagship, cheaper than mid-tier" formula that Anthropic has repeated in every Sonnet generation since.
Can you still use Claude 3.5 Sonnet in 2026?
No. Anthropic marked the model as legacy and shut it down — Claude 3.5 Sonnet is no longer available in the API or on partner platforms. The intermediate versions met the same fate: Claude Sonnet 4, for example, was retired on June 15, 2026.
That's the normal model lifecycle: each new Sonnet generation is smarter, faster, and better value than the last, so there's no reason to keep the old ones in production. Everything 3.5 could do, the current Sonnets do better — from code to image analysis.
Claude Sonnet generations: from 3.5 to Sonnet 5
| Version | Release | What it added | Status |
|---|---|---|---|
| Claude 3.5 Sonnet | June 2024 | leap in coding and vision, Artifacts | retired |
| Claude 3.7 Sonnet | February 2025 | hybrid reasoning, SWE-bench 68% | retired |
| Claude Sonnet 4 | May 2025 | agentic coding, long action chains | retired 15 Jun 2026 |
| Claude Sonnet 4.5 | September 2025 | SWE-bench 77.2%, thinking budget | outdated |
| Claude Sonnet 4.6 | February 2026 | adaptive thinking, 1M-token context | available in GPTunneL |
| Claude Sonnet 5 | summer 2026 | agency, answer self-checking, 1M context | current version |
For a closer look at the milestones, see our reviews: Claude 3.7 Sonnet review, Claude Sonnet 4.6 vs Opus 4.6, and the Claude Sonnet 5 review.
What to use instead of Claude 3.5 Sonnet
The latest Claude Sonnet is Claude Sonnet 5 — the direct heir to the 3.5 philosophy: the balance of intelligence, speed, and price, at 2026 level. It plans multi-step tasks on its own, checks its answers before responding, handles up to 1 million tokens of context, and has become the default model for most Claude users.
Both current Sonnets are available in GPTunneL:
- Claude Sonnet 5 — for everything: code, writing, document analysis, agentic workflows.
- Claude Sonnet 4.6 — if your workflows and prompts are tuned for the 4.x generation.
If you're torn between the model lines, here's the rule of thumb: Sonnet is the everyday workhorse, Opus is maximum depth for tasks where mistakes are expensive. The whole family lives on the Claude page in GPTunneL.
How much does Claude Sonnet cost?
No subscription needed: in GPTunneL you pay only for the tokens you use. Claude Sonnet 5 costs $0.004 per 1K input tokens and $0.02 per 1K output tokens — a typical conversation costs next to nothing. See the pricing page for exact numbers for your workload.
FAQ
What is the latest Claude Sonnet version? Claude Sonnet 5, released in summer 2026. From the previous generation, Claude Sonnet 4.6 with its 1M-token context is still in service.
Does Claude 3.5 Sonnet still work anywhere? No. Anthropic shut the model down; it has been removed from the API and partner catalogs. The current replacement is Claude Sonnet 5.
What's the difference between Sonnet and Opus? Sonnet is the sweet spot of price and speed for everyday tasks; Opus is the flagship for the hardest reasoning. See our detailed Sonnet 4.6 vs Opus 4.6 comparison.
Can I download Claude Sonnet and run it locally? No, Anthropic's model weights are closed. Claude Sonnet runs only in the cloud: on claude.ai with a subscription, or in GPTunneL with pay-per-token pricing.
Is there a free tier? No free tier — but there's no subscription either. In GPTunneL you top up your balance and pay only for the tokens you actually use.
Bottom line
Claude Sonnet 3.5 did its job — and did it brilliantly — but in 2026 there's no reason to look for it. Try Claude Sonnet 5 in GPTunneL — no VPN needed, no subscription, pay per use: open a chat with the model and run your own tasks through it, from code to long documents. Everything people loved 3.5 for simply works better here.