1M token context
About a million tokens in one conversation, enough for a large repository or a stack of long documents.
Anthropic released Claude Opus 5 on July 24, 2026: near the intelligence of Claude Fable 5 at half the input price, with a 1 million token context window, vision, and 96.0% on SWE-bench Verified. On avots.ai you run it pay per use, with no Claude Pro or Max plan, and the same balance also covers GPT-5.6, Gemini, Kimi K3 and 40+ image, video and music models.
About a million tokens in one conversation, enough for a large repository or a stack of long documents.
Per million tokens, the same rate as Opus 4.8 and half the input price of Claude Fable 5.
Real GitHub issues fixed end to end, with 93.2% on GPQA Diamond for graduate level science.
Send screenshots, diagrams, charts and scanned pages alongside your text prompt.
Low through max effort settings trade cost against depth, so cheap tasks stay cheap.
The same model at roughly 2.5 times the output speed for 10 USD in and 50 USD out.
| Claude Opus 5 | Claude Fable 5 | GPT-5.6 Sol | |
|---|---|---|---|
| Price per 1M in / out | $5 / $25 | $10 / $50 | $5 / $30 |
| Context window | 1M tokens | 1M tokens | 1.05M tokens |
| Frontier-Bench v0.1 | 43.3% | 33.7% | not published |
| Vision | yes | yes | yes |
| Released | July 24, 2026 | June 9, 2026 | July 9, 2026 |
| On avots.ai | pay per use | pay per use | pay per use |
Opus 5 more than doubles Opus 4.8's Frontier-Bench score and passes Fable 5 on that agentic coding evaluation while costing half as much per input token. Switch between all three in the picker, one balance covers them.
43.3% on Frontier-Bench v0.1 for terminal coding agents, ahead of Claude Fable 5 at 33.7%.
Opus 5 verifies its output and retries until the task passes, which is what long agent runs need.
70.6% on OSWorld 2.0 for driving a real desktop, and 26.0% on AutomationBench for business workflows.
30.2% on ARC-AGI-3, about three times the next best model reported at launch, plus 89.1% Terminal-Bench 2.1.
Sign up on app.avots.ai with email, Google or Telegram. Takes a minute.
1,000 tokens cost 0.99 EUR. No plan to choose, no renewal, the balance sits there until you use it.
Choose it in the model picker, or let Agent Avots route for you. Every reply shows its token cost.
app.avots.ai on desktop and mobile, no install needed.
Claude Opus 5 in @AvotsAIbot, same balance, from any phone.
Use it inside Claude Desktop, Cursor or Cline via the MCP server.
The OpenAI compatible API takes your existing client, one key for every model.
1,000 tokens cost 0.99 EUR and a short Opus 5 answer costs a handful of them. The same balance unlocks GPT-5.6, Gemini, Kimi K3, Veo video and Nano Banana images, with no Claude subscription.
Sign up and chat ✦Every LLM on avots.ai and which one to pick per task.
Anthropic's most agentic Sonnet, 1M context, near Opus quality for less.
OpenAI's newest family: Sol, Terra and Luna tiers, 1M context.
Moonshot's frontier open model, number one at coding, 1M context.
What to use instead of ChatGPT Plus and how one balance replaces it.
Google against OpenAI, task by task, tested on one balance.
avots.ai runs 40+ AI models for chat, image, video and audio on one balance. See the full list on the Tools page or open the app and start creating.