Claude Opus 5 vs Qwen3.8 2.4T A95B
API pricing, context, and independently measured performance, side by side. Qwen3.8 2.4T A95B starts 60% cheaper on input than Claude Opus 5 at list prices.
Claude Opus 5
Anthropic · released Jul 24, 2026
Best input$5.00 / MTok
Best output$25.00 / MTok
Context1M
Providers9
Fastest measured2s° TTFT · 131 tok/s
VisionToolsReasoningStructured outputPrompt caching
Qwen3.8 2.4T A95B
Qwen · released Aug 12, 2026
Best input$2.00 / MTok
Best output$6.00 / MTok
Context1.05M
Providers6
Fastest measured1s° TTFT · 122 tok/s
Open weightsToolsReasoningStructured outputPrompt caching
Cheapest providers, measuredUSD / MTOK
| MODEL | PROVIDER | INPUT | OUTPUT | TTFT | TPS | SMOKE |
|---|---|---|---|---|---|---|
| Claude Opus 5 | CPOClaude Platform on AWS | $5.00 | $25.00 | 4.7s° | 442° tok/s | — |
| Claude Opus 5 |  Azure | $5.00 | $25.00 | 2s° | 131° tok/s | — |
| Claude Opus 5 |  Anthropic | $5.00 | $25.00 | 4.2s | 205 tok/s | — |
| Qwen3.8 2.4T A95B |  DeepInfra | $2.00 | $6.00 | 1.3s° | 76° tok/s | 4/5 |
| Qwen3.8 2.4T A95B |  SiliconFlow | $2.00 | $6.00 | 2.7s° | 49° tok/s | 5/5 |
| Qwen3.8 2.4T A95B | MOModal | $2.00 | $6.00 | 1s° | 122° tok/s | 4/5 |
TTFT/TPS: ModelIndex-measured medians (P50, trailing 72h) · chat10k profile (~10K in / 200 out) · ° = measured via OpenRouter routing (adds a hop) · SMOKE: of 5 programmatic capability probes passed · methodology v1.1 — full details on each model page