Observed 2026-06-03 11:12. Prices are USD per 1M tokens (from OpenRouter public pricing). Input: Claude Opus 4.8 $5 / DeepSeek R1 $0.7, output: $25 / $2.5.
✓ = the mechanically better side per row (cheaper price / longer context). Magnitude only, not a rating.
| Field | Claude Opus 4.8 anthropic/claude-opus-4.8 | DeepSeek R1 deepseek/deepseek-r1 |
|---|---|---|
| Input priceUSD/1M | $5 | $0.7✓ |
| Output priceUSD/1M | $25 | $2.5✓ |
| Cache readUSD/1M | $0.5 | Accumulating… |
| Context windowtokens | 1,000k✓ | 163.84k |
Source: OpenRouter /api/v1/models → collect_prices.py → Source(price)。 · Claude Opus 4.8: anthropic/claude-opus-4.8 · DeepSeek R1: deepseek/deepseek-r1
Aggregated observations, not predictions. Prices change (observation time noted). © TraceBottle for AI
As observed on 2026-06-03 11:12, input is Claude Opus 4.8 $5 / DeepSeek R1 $0.7 per 1M tokens — DeepSeek R1 is cheaper. Output is $25 / $2.5 — DeepSeek R1 is cheaper.
Claude Opus 4.8: 1,000,000 tokens, DeepSeek R1: 163,840 tokens (observed 2026-06-03 11:12).
Observed 2026-06-03 11:12 (UTC). Source: OpenRouter public pricing (/api/v1/models). Prices change — check each provider's official pricing for the latest. This site publishes only observed values, never predictions.