WizardLM-2 8x22B
WizardLM-2 8x22B is a text-focused AI model from Microsoft with an impressive context length of 65,535 tokens and a maximum completion of 8,000 tokens. It excels in handling textual inputs but lacks support for tools, reasoning, or structured output formats. Given its price point at $0.62 per million input and output tokens, it is most suitable for users requiring high-context text processing, such as complex document analysis or creative writing projects. While the lack of benchmark coverage means its performance relative to other models remains unproven, WizardLM-2 8x22B stands out for those prioritizing text-based tasks at a competitive cost.
- Model ID
- microsoft/wizardlm-2-8x22b
- Vendor
- microsoft
- Released
- April 2024
- Tokenizer
- Mistral
- Input Modalities
- text
- Output Modalities
- text
- Max Output
- 8,000 tokens
- Tool Calling
- not supported
- Structured Output
- ✓ supported
- Reasoning Mode
- not supported
- Vision
- text only
- Audio
- no
- Moderated
- no
What it costs in practice
Computed from the current $0.62/M input and $0.62/M output rates. Run your own numbers →
| Job | Tokens | Cost |
|---|---|---|
| Summarize a 50-page report | 30k in / 1.5k out | $0.02 |
| Classify 1,000 customer emails | 500k in / 50k out | $0.34 |
| A month of a busy support chatbot | 5M in / 2M out | $4.34 |
Price & spec history
Tracked daily by PicksByModel since 2026-07-17.
| Date | Input /M | Output /M | Context |
|---|---|---|---|
| 2026-07-22 | $0.62 | $0.62 | 65,535 |
| 2026-07-17 | $0.62 | $0.62 | 65,536 |
Similar models
Quick answers
- How much does WizardLM-2 8x22B cost?
- $0.62 per million input tokens and $0.62 per million output tokens.
- What is WizardLM-2 8x22B's context window?
- 65,535 tokens, roughly 98 pages of text in a single request.
- Does WizardLM-2 8x22B support tool calling?
- It supports structured output.
- Can WizardLM-2 8x22B process images?
- No, it is text-only on the input side.