WizardLM 2 8x22B
WizardLM 2 8x22BCompare inon
WizardLM 2 8x22B measured on 3 of 104 Chat & Knowledge benchmarks3
BenchmarkByScoreRank
Legacy · saturated or no longer reported2
No benchmark matches
Alternatives?
ModelScoreCostPer taskWhat changes
CheaperGPT-5.6 Lunamax×2.1-6%$0.18
OpenKimi K3max×2.1+1,150%$2.00Sampling parameters are fixed+3CheaperGPT-6 Lunamax×2.1-58%$0.068Function calling only with none · Chat Completions+3
CheaperGLM 5.3 FlashmaxOpen×2.1-51%$0.28Defaults to max · Self-hosted+4
CheaperClaude Haiku 5.5max×2.1-58%$0.21Remove sampling parameters+5ModelScoreCostPer taskWhat changes
GPT-5.6 Lunamax×2.1-6%$0.18GPT-6 Lunamax×2.1-58%$0.068Function calling only with none · Chat Completions+3GLM 5.3 FlashmaxOpen×2.1-51%$0.28Defaults to max · Self-hosted+4Claude Haiku 5.5max×2.1-58%$0.21Remove sampling parameters+5
Qwen 3.5 Plusunstated+95%-46%$0.12 / $0.69ModelScoreCostPer taskWhat changes
GPT-5.6 Lunalow+90%-6%$0.01GPT OSS 20BhighOpen+6%-93%$0.011
Gemini 3.1 Flash Litehigh+88%+17%$0.06GPT-6 Lunamax×2.1-58%$0.068Function calling only with none · Chat Completions+3GPT OSS 120BhighOpen+74%-88%$0.092ModelScoreCostPer taskWhat changes
Kimi K3maxOpen×2.1+1,150%$2.00Sampling parameters are fixed+3GLM 5.2maxOpen×2.1+348%$1.47
DeepSeek V4 Pro 0813maxOpen×2.1+313%$0.67Pass back reasoning_content · Chat Completions+4
MiniMax M3unstatedOpen×2.1+9%$0.30 / $1.20No strict JSON+1GLM 5.3maxOpen×2.1+348%$2.01Reasoning cannot be disabled+3ModelScoreCostPer taskWhat changes
Phi 4unstatedOpen+29%—$0.056 / $0.11Context: 66K → 16KPhi 3 Medium 128K InstructOpen-36%——Phi 4 Reasoning PlusOpenn/a—$0.07 / $0.35Context: 66K → 33KPhi 4 ReasoningOpenn/a——Context: 66K → 33KPhi 3.5 Mini InstructOpenn/a——Phi 3 Medium 4K InstructOpenn/a—$0.14 / $0.14Context: 66K → 4KPhi 3 Small 8K InstructOpenn/a——Context: 66K → 8KPhi 3 Mini 128K InstructOpenn/a——All Microsoft models →Benchmarks
in Chat & KnowledgeMMLU-Pro · official board39.2#24
GPQA Diamond · Epoch43.4#110
MATH Level 525.7#27
FAQ
How much does WizardLM 2 8x22B cost?›
$0.48 per 1M input tokens and $0.48 per 1M output tokens (cheapest host: DeepInfra).
Is WizardLM 2 8x22B free to use?›
There is no free API; the cheapest paid price is $0.48 per 1M input tokens.
How good is WizardLM 2 8x22B, and what is it best at?›
#110 of 133 in Chat & Knowledge.
Is WizardLM 2 8x22B better than Llama 3.1 70B Instruct?›
On GPQA Diamond · Epoch, WizardLM 2 8x22B scores 43.4 vs 44.2 — about the same.
How can I use WizardLM 2 8x22B?›
Through the API of DeepInfra, Novita.
What is the context window of WizardLM 2 8x22B?›
66K tokens.
What is WizardLM 2 8x22B?›
WizardLM 2 8x22B is a language model from Microsoft.
What are the best alternatives to WizardLM 2 8x22B?›
GPT-5.6 Luna (cheaper), Kimi K3 (open weights).