Best Local & Open-Source Models for OpenClaw (2026)

Community-voted LLM rankings for running OpenClaw with local and open-source models

Discussion · OpenClaw

After Claude ban I found my new main model

I've been using OpenClaw for months with only Opus 4.6, Sonnet 4.6, and GPT 5.3/5.4. I'm the kind of person who needs the flagship model as long as budgets are reasonable. Claude is dead. OpenAI made business plan quotas unusable. So I went shopping for alternatives. GLM 5.1 and 5 Turbo: absolute garbage for agentic tasks and automation. Couldn't even write a simple Reddit reply without flooding Telegram with code dumps. Felt like talking to a drunk model. Cancelled. (They said "we'll refund" — still waiting 3 weeks later.) MiMo V2 Pro: using it since launch, really liked it. Honestly got Opus/GPT vibes in many ways. After Claude banned OpenClaw yesterday I got the Token Plan (standard $16). Terrible credit system. Everything in OpenClaw deducts from credits. Session history, bootstrap MD content, tool outputs, cache — literally everything. One month's quota gone in 1 day after filling just 2 session contexts. Horribly inefficient. I will never pay again until they fix the credit logic. Kimi: reviews were bad, never tried. Grok: community feedback looked bad, skipped it. Gemini: no monthly payment option. If there was I'd probably use it, but it's too expensive. So I went with the most popular alternative: Minimax 2.7. When MiMo and GPT failed to handle my nit cron task and Minimax M2.7 solved it in 5 minutes. And the quota on Minimax is impossible to exhaust. — I was shocked. How are they this generous? If anyone knows please explain because it really feels like it won't run out. Tested browser automations. It's not as smart as Opus, but for my automation tasks, light coding work, and being a personal agent — it's enough. If it falls short, I might rotate between a few GPT Plus subscriptions for GPT 5.4 access. Right now, price/performance-wise, Minimax and GPT Plus x 2 accounts are the only efficient OpenClaw model options I could find.

6 comments·Join the discussion →

5 Tips to Save Money on OpenClaw

All Models 💰 Cheapest Models

Provider	Model	Input $/M	Output $/M	Score
MI Mistral AI	Mixtral 8x7B Instruct	$0.140	$0.420	71+6
MI Mistral AI	Mistral 7B Instruct v0.2	$0.140	$0.200	40+4
MI Mistral AI	Mixtral 8x7B Instruct (HF)	$0.500	$0.500	20+2
QW Qwen	Qwen2.5 Coder 7B	$0.200	$0.200	10+1
QW Qwen	Qwen3.5 0.8B	$0.000	$0.000	000
QW Qwen	Qwen3.5 2B (Non-reasoning)	$0.000	$0.000	000
QW Qwen	Qwen3.5 4B (Non-reasoning)	$0.000	$0.000	000
G Google	Gemma 3 270M Instruct	$0.000	$0.000	000
M Meta-llama	Code Llama 7B Instruct	$0.200	$0.200	000
MI Mistral AI	Snorkel Mistral PairRM DPO	$0.200	$0.200	000
QW Qwen	Qwen3 8B	$0.000	$0.000	000
FW Fireworks	Marin 8B Instruct	$0.000	$0.000	000
LQ Liquid	LiquidAI/LFM2-8B-A1B	$0.010	$0.020	000
LQ Liquid	LiquidAI/LFM2-2.6B	$0.010	$0.020	000
G Google	CodeGemma 2B	$0.100	$0.100	000
G Google	CodeGemma 7B	$0.200	$0.200	000
NO Nousresearch	Chronos Hermes 13B v2	$0.200	$0.200	000
BY Bytedance	UI-TARS 7B	$0.100	$0.200	000
-	GLM 4.1V 9B Thinking	$0.000	$0.000	000
AP Alfredpros	CodeLLaMa 7B Instruct Solidity	$0.800	$1.200	000
CO Cohere	Command R7B (12-2024)	$0.037	$0.150	000
E Essential AI	Rnj 1 Instruct	$0.150	$0.150	000
QW Qwen	Qwen2 VL 7B Instruct	$0.200	$0.200	000
A Allenai	Molmo 2 4B	$0.200	$0.200	000
D Deepcogito	Cogito v1 Preview Llama 3B	$0.100	$0.100	000
EL Eleutherai	Pythia 12B	$0.200	$0.200	000
M Meta-llama	Code Llama 13B	$0.200	$0.200	000
M Meta-llama	Code Llama 13B Instruct	$0.200	$0.200	000
M Meta-llama	Code Llama 13B Python	$0.200	$0.200	000
M Meta-llama	Code Llama 7B	$0.200	$0.200	000
MI Mistral AI	Mistral 7B OpenOrca	$0.200	$0.200	000
QW Qwen	Qwen2 7B Instruct	$0.000	$0.000	000
QW Qwen	Qwen2 VL 2B Instruct	$0.100	$0.100	000
QW Qwen	Qwen3 4B	$0.200	$0.200	000
NO Nousresearch	OpenHermes 2.5 Mistral 7B	$0.170	$0.170	000
A Allenai	Molmo 2 8B	$0.200	$0.200	000
QW Qwen	CodeQwen 1.5 7B	$0.200	$0.200	000
QW Qwen	Qwen2.5 1.5B Instruct	$0.000	$0.000	000
QW Qwen	Qwen2.5 7B	$0.200	$0.200	000
QW Qwen	Qwen2.5 Coder 0.5B	$0.200	$0.200	000
QW Qwen	Qwen2.5 Coder 0.5B Instruct	$0.200	$0.200	000
QW Qwen	Qwen2.5 Coder 1.5B	$0.200	$0.200	000
QW Qwen	Qwen2.5 Coder 3B Instruct	$0.000	$0.000	000
QW Qwen	Qwen3 0.6B	$0.000	$0.000	000
QW Qwen	Qwen3 4B Instruct 2507	$0.000	$0.000	000
D Deepcogito	Cogito v1 Preview Llama 8B	$0.000	$0.000	000
DS Deepseek	DeepSeek Coder 1.3B Base	$0.100	$0.100	000
DS Deepseek	DeepSeek Coder 7B Base	$0.200	$0.200	000
NO Nousresearch	Toppy M 7B	$0.200	$0.200	000
NO Nousresearch	Zephyr 7B Beta	$0.200	$0.200	000
AR Arcee AI	Trinity Mini	$0.045	$0.150	000
A Allenai	Olmo 3 7B Instruct	$0.000	$0.000	000
A Allenai	Olmo 3 7B Think	$0.000	$0.000	000
DS Deepseek	DeepSeek Coder 7B Base v1.5	$0.200	$0.200	000
DS Deepseek	DeepSeek Coder 7B Instruct v1.5	$0.200	$0.200	000
DS Deepseek	R1 Distill Qwen 1.5B	$0.180	$0.180	000
G Google	Gemma 2B	$0.100	$0.100	000
G Google	Gemma 7B	$0.200	$0.200	000
M Meta-llama	Llama 2 13B	$0.200	$0.200	000
M Meta-llama	Llama 2 13B Chat	$0.200	$0.200	000
NO Nousresearch	Hermes 2 Pro Mistral 7B	$0.200	$0.200	000
OG Opengvlab	InternVL3 8B	$0.200	$0.200	000
QW Qwen	Qwen2.5 Coder 1.5B Instruct	$0.200	$0.200	000
QW Qwen	Qwen2.5 Coder 3B	$0.100	$0.100	000
QW Qwen	Qwen2.5 VL 3B Instruct	$0.100	$0.100	000
QW Qwen	Qwen3 1.7B	$0.000	$0.000	000
MI Mistral AI	Ministral 3 8B 2512	$0.150	$0.150	000
MI Mistral AI	Ministral 3 3B 2512	$0.100	$0.100	000
NV Nvidia	Nemotron Nano 12B 2 VL	$0.000	$0.000	000
NV Nvidia	Nemotron Nano 9B V2	$0.000	$0.000	000
NO Nousresearch	Nous Capybara 7B v1.9	$0.200	$0.200	000
MI Mistral AI	Mistral 7B v0.2	$0.200	$0.200	000
M Meta-llama	Llama 2 7B Chat	$0.000	$0.000	000
M Meta-llama	Llama 3.2 1B	$0.100	$0.100	000
M Meta-llama	Llama 3.2 3B	$0.100	$0.100	000
M Meta-llama	Llama 3 8B	$0.050	$0.080	000
MI Mistral AI	Mistral 7B	$0.200	$0.200	000
NO Nousresearch	Nous Hermes 2 Mixtral 8x7B DPO	$0.500	$0.500	000
NO Nousresearch	Nous Hermes Llama 2 13B	$0.170	$0.170	000
NO Nousresearch	Nous Hermes Llama 2 7B	$0.200	$0.200	000
NO Nousresearch	OpenHermes 2 Mistral 7B	$0.200	$0.200	000
NV Nvidia	Nemotron Nano 12B V2	$0.200	$0.200	000
Z Z-ai	GLM 4.5V	$0.000	$0.000	000
DS Deepseek	DeepSeek R1 0528 Qwen3 8B	$0.200	$0.200	000
M Meta-llama	Llama 2 7B	$0.200	$0.200	000
CG Cognitive Computations	Dolphin 2.6 Mixtral 8x7B	$0.500	$0.500	000
G Google	Gemma 7B Instruct	$0.070	$0.070	000
M Meta-llama	Llama 3 8B Instruct (HF)	$0.200	$0.200	000
MI Mistral AI	Mixtral 8x7B	$0.240	$0.240	000
NO Nousresearch	OpenChat 3.5 0106	$0.200	$0.200	000
U Upstage	SOLAR 10.7B Instruct v1	$0.300	$0.300	000
QW Qwen	Qwen2.5 Coder 7B Instruct	$0.200	$0.200	000
EL Eleutherai	Llemma 7b	$0.000	$0.000	000
G Google	Gemma 3 4B	$0.000	$0.000	000
G Google	Gemma 3 12B	$0.040	$0.130	000
MS Microsoft	Phi 4 Multimodal Instruct	$0.000	$0.000	000
A Allenai	OLMo 7B Instruct	$0.200	$0.200	000
A Austism	Austism/chronos-hermes-13b	$0.300	$0.300	000
G Google	Gemma 2B	$0.100	$0.100	000
IB Ibm	ALLaM 7B Instruct Preview	$0.000	$0.000	000
L Lmsys	lmsys/vicuna-13b-v1.5	$0.300	$0.300	000
NO Nousresearch	OpenChat 7B	$0.060	$0.060	000
QW Qwen	Qwen1.5 0.5B	$0.100	$0.100	000
QW Qwen	Qwen1.5 0.5B Chat	$0.100	$0.100	000
TD Thedrummer	UnslopNemo 12B	$0.400	$0.400	000
MI Mistral AI	Ministral 8B	$0.100	$0.100	000
MI Mistral AI	Ministral 3B	$0.040	$0.040	000
QW Qwen	Qwen2.5 7B Instruct	$0.040	$0.100	000
TD Thedrummer	Rocinante 12B	$0.170	$0.430	000
M Meta-llama	Llama 3.2 3B Instruct	$0.000	$0.000	000
M Meta-llama	Llama 3.2 1B Instruct	$0.000	$0.000	000
M Meta-llama	Llama 3.2 11B Vision Instruct	$0.000	$0.000	000
QW Qwen	Qwen3 VL 8B Instruct	$0.080	$0.200	000
NS Neversleep	Lumimaid v0.2 8B	$0.000	$0.000	000
MI Mistral AI	Pixtral 12B	$0.100	$0.100	000
QW Qwen	Qwen2.5-VL 7B Instruct	$0.200	$0.200	000
MS Microsoft	Phi-3.5 Mini 128K Instruct	$0.000	$0.000	000
S1 Sao10k	Llama 3 8B Lunaris	$0.040	$0.050	000
MI Mistral AI	Mistral Nemo	$0.000	$0.000	000
NO Nousresearch	Hermes 2 Pro - Llama-3 8B	$0.140	$0.140	000
G Google	Gemma 2 9B	$0.000	$0.000	000
MI Mistral AI	Mistral 7B Instruct	$0.059	$0.059	000
MI Mistral AI	Mistral 7B Instruct v0.3	$0.140	$0.200	000
MS Microsoft	Phi-3 Mini 128K Instruct	$0.000	$0.000	000
FW Fireworks	Sarvam M	$0.000	$0.000	000
MI Mistral AI	Mistral Tiny	$0.140	$0.420	000
M Meta-llama	Llama 3 8B Instruct	$0.030	$0.040	000
U9 Undi95	ReMM SLERP 13B	$0.300	$0.300	000
GR Gryphe	MythoMax 13B	$0.060	$0.060	000
G Google	Gemma 3 1B	$0.000	$0.000	000
QW Qwen	Qwen2 1.5B	$0.000	$0.000	000
QW Qwen	Qwen2 1.5B Instruct	$0.020	$0.020	000
QW Qwen	Qwen3 0.6B Base	$0.000	$0.000	000
QW Qwen	Qwen3 1.7B Base	$0.000	$0.000	000
QW Qwen	Qwen3 4B Base	$0.000	$0.000	000
QW Qwen	Qwen3 8B Base	$0.000	$0.000	000
O Open Orca	Open-Orca/Mistral-7B-OpenOrca	$0.200	$0.200	000
QW Qwen	Qwen2.5 0.5B Instruct	$0.200	$0.200	000
G Google	Gemma 3n 4B	$0.060	$0.120	01-1
QW Qwen	Qwen3.5 9B	$0.100	$0.150	01-1
M Meta-llama	Llama 3.1 8B Instruct	$0.020	$0.050	01-1

Vote for open-source models that work well (or don't) with OpenClaw.

Pricing from OpenRouter.

Running OpenClaw with a Local Model or LLM

OpenClaw supports local models via OpenAI-compatible APIs. The easiest way to run a model locally is with Ollama or llama.cpp, both of which expose a local API endpoint that OpenClaw can connect to.

VRAM Requirements

7B models need ~6GB VRAM (4-bit quantized). 13B models need ~10GB. 70B models need ~40GB or multi-GPU. If you're on CPU-only, smaller quantized models (Q4) are usable but slow.

Recommended: Ollama

Install Ollama, run ollama pull llama3.3, then point OpenClaw to http://localhost:11434/v1.

Alternative: llama.cpp server

For more control over quantization and performance, llama.cpp's server mode gives you a full OpenAI-compatible API with fine-grained settings.

Compare all local LLM runners →

About This Leaderboard

This leaderboard shows community votes specifically for open-source and locally-runnable models and LLMs used with OpenClaw. Models are filtered to include those from open-source providers: Meta (Llama), Mistral, Qwen (Alibaba), Google (Gemma), DeepSeek, Microsoft (Phi), and similar open-weight families.

Running local models with OpenClaw gives you full privacy, zero API costs, and offline capability — at the cost of needing hardware and accepting some quality trade-offs versus frontier API models.

Best Local & Open-Source Models for OpenClaw (2026)

After Claude ban I found my new main model

Running OpenClaw with a Local Model or LLM

About This Leaderboard

Frequently Asked Questions

Tools

Directories

Models & Pricing

Endpoints

Rankings

News