Anthropic: Claude Opus 4.8
Providers
Where this model runs, and what each route costs. Requests are sent to a healthy provider automatically; if one errors, the gateway retries against another serving the same model.
| Provider | Input /M | Output /M | Cache read /M | Latency | Throughput |
|---|---|---|---|---|---|
| Anthropic | $2.00 | $10.00 | $0.500 | 7.99s | 70 tps |
Capabilities (18)
Supported parameters (7)
Any other parameter you send is ignored rather than rejected.
Pricing
What this model costs to run, next to the rate it is posted at. Caching and discounts mean the price actually paid is often below the listed one.
| Rate | Listed | Charged | Unit |
|---|---|---|---|
| Input | $5.00 | $2.00 | per 1M tokens |
| Output | $25.00 | $10.00 | per 1M tokens |
| Cache read | $0.500 | $0.500 | per 1M tokens |
| Cache write | $6.25 | $6.25 | per 1M tokens |
| Cache write (1h) | $10.00 | $10.00 | per 1M tokens |
Performance
Throughput is how fast the model writes (tokens per second — higher is better). Latency is total round-trip time (lower is better). TTFT is time-to-first-token — how long before you see anything appear (lower is better). Throughput and TTFT are measured on streaming requests, which are the only ones with a first-token moment to time.
Throughput
Latency
Time to first token
Cache hit rate
Availability
The share of requests to this model that completed successfully. When an upstream provider errors, the gateway retries against another provider serving the same model, so a single provider incident does not necessarily show up here.
Days are UTC days. Days with no requests are omitted rather than drawn at 100% — no traffic is not evidence of availability.
Apps
Public apps that send the most traffic to this model. Good signal for what real production workloads look like — and a hint at which use cases this model is best suited for.
- 1.Oxyy Tester73.9K tokens
Activity
Token volume and request traffic to this model over time. Daily totals, UTC.
Prompt tokens measure input size. Reasoning tokens show internal thinking before a response. Completion tokens reflect total output length.