OpenAI: gpt-oss-20b vs Anthropic: Claude Haiku 4.5
Head-to-head API pricing and cost comparison between OpenAI’s OpenAI: gpt-oss-20b and Anthropic’s Anthropic: Claude Haiku 4.5. Prices auto-refresh daily from OpenRouter.
OpenAI: gpt-oss-20b is 97% cheaper for input tokens; OpenAI: gpt-oss-20b also wins on output tokens.
Side-by-side comparison
| Spec | OpenAI: gpt-oss-20b | Anthropic: Claude Haiku 4.5 |
|---|---|---|
| Input price (per 1M) | $0.03 | $1.00 |
| Cached input (per 1M) | — | $0.10 |
| Output price (per 1M) | $0.14 | $5.00 |
| Batch input (per 1M) | $0.01 | $0.50 |
| Batch output (per 1M) | $0.07 | $2.50 |
| Reasoning price (per 1M) | — | — |
| Context window | 131K | 200K |
| Vision support | No | Yes |
| Caching support | No | Yes |
| Batch API | Yes | Yes |
| Reasoning capability | No | No |
Monthly cost at volume
Estimated monthly API spend at common production traffic levels (input/output tokens per request shown).
| Volume | OpenAI: gpt-oss-20b | Anthropic: Claude Haiku 4.5 | Savings |
|---|---|---|---|
1K req/day 500in / 200out tokens | $1.29 | $45.00 | $43.71 OpenAI: gpt-oss-20b wins |
10K req/day 1500in / 500out tokens | $34.50 | $1,200 | $1,166 OpenAI: gpt-oss-20b wins |
100K req/day 3000in / 800out tokens | $606.00 | $21,000 | $20,394 OpenAI: gpt-oss-20b wins |
1M req/day 8000in / 2000out tokens | $15,600 | $540,000 | $524,400 OpenAI: gpt-oss-20b wins |
Adjust input/output token counts, request volume, batch & cached pricing.
Related comparisons
Frequently asked questions
Which is cheaper, OpenAI: gpt-oss-20b or Anthropic: Claude Haiku 4.5?
For input tokens, OpenAI: gpt-oss-20b is roughly 97% cheaper at $0.03/1M vs $1.00/1M. For output tokens, OpenAI: gpt-oss-20b wins at $0.14/1M. Real-world cost depends on your input/output ratio — use the calculator to model your actual workload.
What’s the context window difference?
OpenAI: gpt-oss-20b has a context window of 131K tokens. Anthropic: Claude Haiku 4.5 offers 200K tokens. Larger context windows are valuable for long documents, RAG pipelines, and multi-turn conversations — but they come with higher input-token bills if you fill them every request.
Should I use OpenAI: gpt-oss-20b or Anthropic: Claude Haiku 4.5?
Choose OpenAI: gpt-oss-20b if you’re already on the OpenAI stack, want broad ecosystem support, or prefer its lower input price. Choose Anthropic: Claude Haiku 4.5 for Anthropic’s ecosystem, native vision input, or its differentiated capabilities. Run a small benchmark on your own prompts before committing — price is only one axis.
How are these prices kept current?
Prices are pulled directly from OpenRouter’s public models API once every 24 hours via a Convex cron job, then normalized to per-1M-token figures. Last refresh: Apr 21, 2026.