WizardLM-2 8x22B
WizardLM-2 8x22B is a previous Microsoft Wizard snapshot, not a 2026 default open chat model. Hugging Face still has community mirrors such as lucyknada/microsoft_WizardLM-2-8x22B.
Compare leftover WizardLM-2 7B if you wanted the leftover smaller twin already rewritten as previous, leftover Mistral 7B Instruct if you wanted another leftover Mixtral-era chat path, and Gemma 4 26B A4B if you wanted a current open path.
Key Features
- Previous Mixtral-era Wizard snapshot: older copy treated 8x22B as Microsoft's current Wizard flagship.
- Not a 2026 census: Do not keep leftover most-advanced / SoTA open claims unless you re-run a first-party card.
- Community mirrors still exist: Confirm Hugging Face before you hard-code a host.
- No invented dollar row: Confirm first-party hosting before you quote a price.
Use Cases
- People with a WizardLM-2 8x22B bookmark who need the leftover latest copy dropped.
- People who were about to paste a SoTA open crown into a 2026 deck.
Limitation: We did not re-run a Wizard bench or invent a 2026 dollar row.
Pricing
First-party pages on 2026-08-18.
| Piece | Price | Notes |
|---|---|---|
| WizardLM-2 8x22B | Previous snapshot | Confirm community Hugging Face mirrors. No invented dollar row. |
| Leftover smaller twin | WizardLM-2 7B | WizardLM-2 7B. |
| Live open path | Gemma 4 | Gemma 4 26B A4B. |
If a leftover blog still calls WizardLM-2 8x22B Microsoft's current Wizard flagship, treat this page as the correction.
Getting Started
- Confirm the Hugging Face mirror before you hard-code WizardLM-2 8x22B.
- Confirm leftover WizardLM-2 7B if you wanted the leftover smaller twin.
- Confirm Gemma 4 26B A4B if you wanted a current open path.
- Do not paste leftover SoTA open copy into a 2026 deck.
Frequently Asked Questions
Is WizardLM-2 8x22B still Microsoft's latest Wizard?
No. Treat it as a previous snapshot.
Same as WizardLM-2 7B?
No. WizardLM-2 7B is the leftover smaller twin already rewritten as previous.
Did we invent a price?
No. Confirm first-party hosting before you quote one.
Alternatives
- WizardLM-2 7B: Leftover smaller Wizard twin.
- Mistral 7B Instruct: Leftover Mixtral-era chat path.
- Gemma 4 26B A4B: Live open path if you wandered here for a current model.
Tips
- Call WizardLM-2 8x22B a previous snapshot, not the current Wizard default.
- Confirm Hugging Face before you hard-code a host.
Conclusion
WizardLM-2 8x22B is a leftover Wizard snapshot, not a 2026 default. Start at the Hugging Face mirror, then decide whether leftover WizardLM-2 7B or live Gemma 4 26B A4B already covers the path you need.
Comments
No comments yet. Be the first to comment!
Related Tools
Related Insights
Grok Bot and Hermes Bot: one person finally gets a think tank and a secretariat
Grok Bot now ships with Cursor Pro+. Hermes Bot runs on a VPS. They are not smarter chat boxes. The think tank advises, the secretariat executes, and you still make the call.
Hook OpenCode Go into Codex on Windows. Do Not Open a Second Toolkit.
A ChatGPT-signed Codex desktop app still shows mostly GPT in the picker. On Windows, enable only OpenCode Go and the Grok, GLM, Kimi, DeepSeek, and MiniMax models you already pay for appear in the same selector. Keys stay local. Native GPT stays put.
What Locks Codex Is the Picker, Not the Models
You already pay for OpenCode Go, Grok, and Z.ai, but the Codex picker still shows mostly GPT. The community project codex-router does not teach another install ritual. It puts subscriptions you already bought back into the selector, keeps keys on the machine, and leaves native GPT alone.