What is the best open source llm right now?

DeepSeek V4 Pro is the current #1 on this list. Rankings move when daily ingest updates SWE-bench, Elo, price, or latency.

How is this list ranked?

Highest reasoning signal among models you can self-host or buy as open weights. Weights: Elo 30%, LiveBench 30%, GPQA 25%, Out $ 15%.

Is preference Elo the only score?

Preference Elo is a crowd vote from LMArena / Arena. People see two hidden answers and pick the one they like more. The model that wins more often gets a higher Elo. That means people preferred it — not that it passed a school test. It is not SWE-bench, not accuracy, and not a number we invent. This list uses a named Stack Engine mix — Elo is only one input when the job is “best chatbot,” not when the job is “cheapest coding agent.”

Why should I trust this over a tweet?

Every cell has a source and an as-of date. We noindex thin pages and we do not fill missing scores. Open the model or pair page and read the snapshot.

Search Intent · best open source llm

Best open-source LLM in 2026

Open-weight models ranked on preference Elo, LiveBench, and SWE-bench. Check the license before you ship.

Weights:Elo 30%LiveBench 30%GPQA 25%Out $ 15%

Current #1 Ranked PickScore 78.0 / 100

DeepSeek V4 Pro

DeepSeek · Open weights

elo p78livebench p82gpqa p81output price per m p65

SWE-bench: 71.6%Elo: 1,536

View full model fact sheet

Complete Ranked Category List

Models ranked by verified benchmark weights across SWE-bench and coding preference evaluations.

#1DeepSeek V4 Pro(DeepSeek)

Score 78.0

eloP78livebenchP82gpqaP81output price per mP65

🏆 #1 Overall Leader in Category

#2DeepSeek V4 Flash(DeepSeek)

Score 69.8

eloP61livebenchP71gpqaP71output price per mP83

Rank #2vs #1

#3Qwen QwQ 32B(Alibaba)

Score 69.2

eloP66livebenchP72gpqaP77output price per mP57

Rank #3vs #1

#4GLM-5.2(Zhipu)

Score 61.6

eloP64livebenchP59gpqaP61output price per mP63

Rank #4vs #1

#5MiniMax M2.5(MiniMax)

Score 53.1

eloP56livebenchP54gpqaP42output price per mP64

Rank #5vs #1

#6Llama 4 Maverick(Meta)

Score 47.1

eloP51livebenchP43gpqaP36output price per mP66

Rank #6vs #1

#7Qwen3 235B(Alibaba)

Score 45.4

eloP39livebenchP47gpqaP49output price per mP49

Rank #7vs #1

#8Qwen 2.5 Coder 32B(Alibaba)

Score 45.1

eloP35livebenchP56gpqaP30output price per mP69

Rank #8vs #1

#9DeepSeek R1(DeepSeek)

Score 36.0

eloP21livebenchP26gpqaP64output price per mP39

Rank #9vs #1

#10Llama 4 Scout(Meta)

Score 35.5

eloP32livebenchP28gpqaP21output price per mP82

Rank #10vs #1

#11Llama 3.1 405B(Meta)

Score 33.3

eloP26livebenchP36gpqaP38output price per mP35

Rank #11vs #1

#12DeepSeek Coder V2(DeepSeek)

Score 32.2

eloP25livebenchP25gpqaP19output price per mP83

Rank #12vs #1

Frequently asked questions

Plain-English methodology and leaderboard answers

: DeepSeek V4 Pro is the current #1 on this list. Rankings move when daily ingest updates SWE-bench, Elo, price, or latency.

What is preference Elo? · How rankings update

Search Intent · best open source llm

Best open-source LLM in 2026

Open-weight models ranked on preference Elo, LiveBench, and SWE-bench. Check the license before you ship.

Weights:Elo 30%LiveBench 30%GPQA 25%Out $ 15%

Current #1 Ranked PickScore 78.0 / 100

DeepSeek V4 Pro

DeepSeek · Open weights

elo p78livebench p82gpqa p81output price per m p65

SWE-bench: 71.6%Elo: 1,536

View full model fact sheet

Complete Ranked Category List

Models ranked by verified benchmark weights across SWE-bench and coding preference evaluations.

#1DeepSeek V4 Pro(DeepSeek)

Score 78.0

eloP78livebenchP82gpqaP81output price per mP65

🏆 #1 Overall Leader in Category

#2DeepSeek V4 Flash(DeepSeek)

Score 69.8

eloP61livebenchP71gpqaP71output price per mP83

Rank #2vs #1

#3Qwen QwQ 32B(Alibaba)

Score 69.2

eloP66livebenchP72gpqaP77output price per mP57

Rank #3vs #1

#4GLM-5.2(Zhipu)

Score 61.6

eloP64livebenchP59gpqaP61output price per mP63

Rank #4vs #1

#5MiniMax M2.5(MiniMax)

Score 53.1

eloP56livebenchP54gpqaP42output price per mP64

Rank #5vs #1

#6Llama 4 Maverick(Meta)

Score 47.1

eloP51livebenchP43gpqaP36output price per mP66

Rank #6vs #1

#7Qwen3 235B(Alibaba)

Score 45.4

eloP39livebenchP47gpqaP49output price per mP49

Rank #7vs #1

#8Qwen 2.5 Coder 32B(Alibaba)

Score 45.1

eloP35livebenchP56gpqaP30output price per mP69

Rank #8vs #1

#9DeepSeek R1(DeepSeek)

Score 36.0

eloP21livebenchP26gpqaP64output price per mP39

Rank #9vs #1

#10Llama 4 Scout(Meta)

Score 35.5

eloP32livebenchP28gpqaP21output price per mP82

Rank #10vs #1

#11Llama 3.1 405B(Meta)

Score 33.3

eloP26livebenchP36gpqaP38output price per mP35

Rank #11vs #1

#12DeepSeek Coder V2(DeepSeek)

Score 32.2

eloP25livebenchP25gpqaP19output price per mP83

Rank #12vs #1

Frequently asked questions

Plain-English methodology and leaderboard answers

: DeepSeek V4 Pro is the current #1 on this list. Rankings move when daily ingest updates SWE-bench, Elo, price, or latency.

What is preference Elo? · How rankings update