whatmodel.sh / WizardLM 70B V1.0
WizardLM 70B V1.0
↗ share
+ contribute
WizardLM 70B V1.0 is an open-weights language model from Microsoft, released August 2023. Its published size is 70B parameters. Licence: llama2.
Catalogue identity notes
Comparison-only. Arena entry 'wizardlm-70b' (text). LMArena's leaderboard table (arena-leaderboard space @6f2eeeb) labels 'wizardlm-70b' as 'WizardLM-70B-v1.0' (Microsoft, Llama 2 Community), linking https://huggingface.co/WizardLM/WizardLM-70B-V1.0. Maker repo WizardLMTeam/WizardLM-70B-V1.0 @54aaeca: licence tag llama2. Parameters: nominal size from the model name. Released 2023-08: month the maker's Hugging Face repo was created (createdAt 2023-08-09). Arena key 'wizardlm-70b'; LMArena's table labels it 'WizardLM-70B-v1.0'.
$ how it compares
The LLMs ranked either side of WizardLM 70B V1.0, out of 828 with a score, then more from Microsoft.
| # | model | score | where | ||
|---|---|---|---|---|---|
| 589 |  Claude Instant 1.2 | 42±1 | comparison only | ||
| 590 |  Llama 3 8B Instruct | 42±0 | local | ||
| 591 | WizardLM 70B V1.0 | 42±1 | local | ||
| 592 |  WizardCoder Python 13B V1.0 | 42±2 | local | ||
| 593 |  Hermes 3 Llama 3.1 8B | 42±1 | local | ||
| more from Microsoft | |||||
| 252 | Phi-4-reasoning-plus 14B | 59±1 | local | ||
| 261 | Phi-4-reasoning 14B | 58±1 | api $0.22/M +local | ||
| 365 | Phi-4 14B | 53±1 | api $0.087/M +local | ||
$ what it's good at
coding▁▁▁▃▆██▇▆▄▃▁ #281 of 458 · 6.1% of its score · 4 tests
Beats Qwen3-VL 32B Instruct, Solar Open 100B (102B-A12B), GLM-4.7-Flash here, though they rank higher overall. Trails DeepSeek LLM 67B Chat here, though they rank lower overall.
| test | result | vs the others in its run | counts |
|---|---|---|---|
| CodeReviewQA: solution identification (hard), invariant accuracy | 49.0% ✎ | ▃··▁▂▂▃▂█▃▄▂ beats 24/47 | 2.4% |
| CodeReviewQA: change localisation (hard), invariant accuracy | 19.2% ✎ | █▂▁▁▂▂▂▃▅▅▅▇ beats 19/47 | 1.9% |
| CodeReviewQA: automated code refinement, exact match rate | 30.9% ✎ | ▁▁▁▁▁▁▂▂▃█▅▄ beats 23/47 | 1.7% |
| CodeReviewQA: change type recognition, invariant accuracy | 78.6% ✎ | ▁▁▁······▁▃█ beats 41/47 | 0.0% |
chat & writing▁▁▁▃▅▄▅█▆▆▇▃ #315 of 393 · 50% of its score · 3 tests
Beats Gemma 4 E4B, OpenThinker3 7B, Gemma 4 E2B here, though they rank higher overall. Trails LFM2 350M, Gemma 3 1B, Flan-T5 XXL here, though they rank lower overall.
| test | result | vs the others in its run | counts |
|---|---|---|---|
| AlpacaEval 2 length-controlled | 17.6% ✎ | ▁▂▂▃▆▆█▆▄▄▁▁ beats 45/122 | 30% |
| Polyrating rating (bias-aware re-fit of Chatbot Arena votes, mid-2024) | 1129.0 ✎ | ▂▁▂▁▃▇█▆▇▄▂▂ beats 62/95 | 10% |
| LMArena (text) | 1184.7 ✎ | ▁▁▁▂▃▃▃▆▆▇█▃ beats 62/379 | 10% |
math▁▂▃▅▆▅▄█▄▁▁▁ #297 of 398 * · 6.3% of its score · 1 test
Beats GPT-4 (0613), Athene-70B, Gemma 2 27B IT here, though they rank higher overall. Trails Gemma 7B (base), DeepSeek R1 Distill Qwen 1.5B, CodeGemma 7B (v1.0) here, though they rank lower overall.
other tests6 not in a task · 37% of its score
| test | result | vs the others in its run | counts |
|---|---|---|---|
| BIG-Bench Hard | 51.1% ✎ | ▁▁▁▁▂▄▅▅█▃▄▁ beats 50/129 | 13% |
| ARC Challenge | 71.9% ✎ | ▆▁▁▂▂▅▇█▄▃▃▂ beats 84/97 | 8.0% |
| MMLU | 55.0% ✎ | ▁▁▁▁▁▂▃▅█▃▆▃ beats 50/183 | 6.2% |
| AGIEval English | 48.7% ✎ | █▄▄··▄··▄█·█ beats 3/9 | 5.1% |
| Open PT-LLM Leaderboard average (9 Portuguese tasks) | 71.1% ✎ | ▃▁▁▁▁▁▁▁▂█▅▁ beats 180/276 | 2.3% |
| ARC Easy | 80.7% ✎ | ▁▂▃▂▃▅▇█▇▆▄▆ beats 62/65 | 2.2% |
$ where to run it
llama.cpp · recommended
version
- Install llama.cpp · Homebrew and winget installs name the commands llama-cli and llama-server instead.macOS / Linux
curl -LsSf https://llama.app/install.sh | shWindowsirm https://llama.app/install.ps1 | iex - Run WizardLM 70B V1.0 · needs ~41.4 GB · from huggingface.co/TheBlokechat in the terminal
llama cli -hf TheBloke/WizardLM-70B-V1.0-GGUF:Q4_K_M - Or start a local server and open the address it prints (http://localhost:9931)
llama serve -hf TheBloke/WizardLM-70B-V1.0-GGUF:Q4_K_M
LM Studio
version
- Install LM Studio · Open the app once before using the lms command.
- Run WizardLM 70B V1.0 · needs ~41.4 GB · from huggingface.co/TheBloke
- Or download it in the terminal
lms get https://huggingface.co/TheBloke/WizardLM-70B-V1.0-GGUF@Q4_K_M -y - Then run lms ls to see its name, and lms chat followed by that name
Ollama
version
- Install Ollama
download it (macOS / Windows) ↗
Linuxcurl -fsSL https://ollama.com/install.sh | sh - Run WizardLM 70B V1.0 · needs ~41.42 GB · from ollama.comchat in the terminal
ollama run wizardlm:70b-llama2-q4_K_M
Weights and documentation: huggingface.co/WizardLMTeam ↗
$ how the score was worked out · Orca 2: Teaching Small Langua… 41%, AlpacaEval 1.0 and 2.0 leader… 30%, Polyrating: A Cost-Effective… 10% and 3 more
Each result implies a score on its own, after adjusting for how hard its test is and who ran it. The score is their weighted average, with a small pull toward the middle for thinly tested models. Each test's weight is its counts column above; here it's added up by who ran the tests. How it works
| source | results | share of the score |
|---|---|---|
 Orca 2: Teaching Small Language Models How to Reason (Mitra et al., Microsoft) | 6 | 41% |
 AlpacaEval 1.0 and 2.0 leaderboards (tatsu-lab/alpaca_eval repository) | 1 | 30% |
| Polyrating: A Cost-Effective and Bias-Aware Rating System for LLM Evaluation (arXiv:2409.00696v3) | 1 | 10% |
 LMArena leaderboard dataset | 1 | 10% |
| CodeReviewQA: The Code Review Comprehension Assessment for Large Language Models (arXiv:2503.16167v2) | 4 | 6% |
 Open Portuguese LLM Leaderboard (archived) | 1 | 2% |
| prior · pull toward the middle | 1% | |
| weighted average | = 41.9 |
$ improve this page
Use a short GitHub form or edit the data through a pull request. Every number needs a source; here's which ones we can use.
data updated · sources and licences · json