Claude Haiku 4.5 vs Qwen3.8 27B
Side-by-side API pricing and Artificial Analysis performance for Claude Haiku 4.5 (Anthropic) and Qwen3.8 27B (Alibaba (Qwen)). Green cells mark the better value in each row.
Prices updated Sep 18, 2026, 7:49 AM UTC · refreshed hourly
| Claude Haiku 4.5 Anthropic | Qwen3.8 27B Alibaba (Qwen) | |
|---|---|---|
| Pricing | ||
| Input $/1M | $1.00 | $0.214 |
| Output $/1M | $5.00 | $2.55 |
| Blended $/1M (3:1) | $2.00 | $0.798 |
| RAG example (30K in / 2K out) | $0.04 | $0.0115 |
| Quality & speed | ||
| Intelligence Index | 17.6 | 33.9 |
| Coding Index | 43.9 | 68.1 |
| Agentic Index | 10.3 | 46.5 |
| Output speed (tok/s) | — | 42 |
| Time to first token | — | 1.22s |
| Specs | ||
| Context window | 200K | 1M |
| Input modalities | text, image, file | text, image, video |
FAQ
Which is cheaper, Claude Haiku 4.5 or Qwen3.8 27B?
Qwen3.8 27B has the lower blended API price at $0.798 per 1M tokens (3:1 input:output mix), versus $2.00 for the other model. Prices are live from OpenRouter and refresh about hourly.
Which is better for RAG workloads?
For a typical RAG request (30K input / 2K output tokens), Qwen3.8 27B costs about $0.0115 per request versus $0.04. Also compare context windows: Claude Haiku 4.5 offers 200K and Qwen3.8 27B offers 1M.
Which scores higher on Artificial Analysis benchmarks?
Qwen3.8 27B leads on the Artificial Analysis Intelligence Index (33.9 vs 17.6). Check coding and agentic indices on this page for workload-specific tradeoffs.