Benchmarks
Frontier list-price board
List prices from the public catalog on 2026-08-16. Latency/success-rate time series belong on /status once live probes are published — this table does not invent unpublished bake-off numbers.
| Model | Input /M | Output /M | Note |
|---|---|---|---|
| Claude Opus 5 claude-opus-5 | $4.75 | $23.75 | ArkLin live |
| Claude Sonnet 5 claude-sonnet-5 | $1.90 | $9.50 | US frontier |
| GPT-5.6 Sol gpt-5.6-sol | $3.80 | $19.00 | US frontier |
| Gemini 3.1 Pro gemini-3.1-pro | $2.00 | $12.00 | US frontier |
| DeepSeek V4 Flash deepseek-v4-flash | $0.14 | $0.28 | CN value |
| Qwen3.8 Max qwen3.8-max | $2.00 | $6.00 | CN frontier — live |
| Qwen3.7 Plus qwen3.7-plus | $0.40 | $1.60 | CN value |
| Kimi K2.7 Code kimi-k2.7-code | $0.95 | $4.00 | CN coding |
| GLM-5.2 glm-5.2 | $1.00 | $4.00 | CN frontier — live |
Method: catalog list prices, 0% token markup. Full catalog: /models · /tools/api-latency-test · /benchmarks/claude-opus · /benchmarks/gpt · /performance