Open LLM Leaderboard

Name: Open LLM Leaderboard - LLM Stats
Creator: LLM Stats
License: https://llm-stats.com/legal/terms-of-service

Ranking the best open LLMs by performance, price, and speed

Select 2-4 models to compare

← Scroll →


MiMo-V2.5-Pro	🇨🇳	Open	1.0M	$0.43	$0.87	121c/s	1,956	28.6	28.7	41.8	–	–	24.9	32.6	2.8	24.7	24.9	20.3	66.7%	—	78.9%	—	—	—	—	—	—	—	—	34.0%	—	—	—	—	—	—	—	—	—	57.2%	1023.2	5.0s	-	Apr. 2026	Xiaomi
GLM-5.1	🇨🇳	Open	200k	$1.40	$4.40	112c/s	1,764	52.3	46.7	42.4	28.9	–	38.8	29.0	–	–	–	–	86.2%	—	—	—	—	—	79.3%	—	—	—	71.8%	52.3%	—	—	40.7%	—	—	—	—	—	—	58.4%	754	12.2s	-	Apr. 2026	ZAI
GLM-5	🇨🇳	Open	200k	$1.00	$3.20	85c/s	1,591	50.3	–	35.5	25.7	–	–	25.8	–	–	–	–	—	—	77.8%	—	—	—	75.9%	—	—	—	67.8%	—	—	—	—	—	—	—	—	—	—	—	744	9.0s	-	Feb. 2026	ZAI
Kimi K2.6	🇨🇳	Open	262.1k	$0.95	$4.00	30c/s	1,553	57.2	50.0	43.2	37.2	–	37.1	31.4	–	–	–	–	90.5%	—	80.2%	—	—	—	86.3%	86.7%	80.1%	—	—	36.4%	—	—	50.0%	—	—	—	—	52.2%	27.9%	58.6%	1000	12.7s	-	Apr. 2026	MoonshotAI
Kimi K2.5	🇨🇳	Open	-	-	-	41c/s	1,462	49.6	47.7	30.5	29.6	–	34.9	13.9	37.0	47.7	47.7	46.7	87.6%	96.1%	76.8%	—	—	—	74.9%	77.5%	78.5%	—	—	50.2%	—	—	—	—	—	—	—	48.7%	—	50.7%	1000	50.1s	-	Jan. 2026	MoonshotAI
DeepSeek-V4-Pro-Max	🇨🇳	Open	1.0M	$1.74	$3.48	27c/s	1,295	55.5	53.3	42.8	32.0	–	33.5	33.2	19.8	40.1	40.3	48.6	90.1%	—	80.6%	—	—	—	83.4%	—	—	—	73.6%	48.2%	57.9%	—	51.8%	—	—	—	—	—	—	55.4%	1600	38.1s	-	Apr. 2026	DeepSeek
Qwen3.5-397B-A17B	🇨🇳	Open	262.1k	$0.60	$3.60	94c/s	1,294	48.4	45.8	29.6	25.4	23.1	27.6	22.7	37.5	53.9	54.0	53.2	88.4%	—	76.4%	—	88.5%	—	69.0%	—	—	—	—	28.7%	—	—	38.3%	—	—	—	—	—	—	—	397	17.9s	-	Feb. 2026	Qwen
Gemma 4 31B	🇺🇸	Open	262.1k	$0.14	$0.40	117c/s	1,284	42.7	39.2	–	–	–	26.9	20.0	24.1	43.0	43.0	37.5	84.3%	—	—	—	88.4%	—	—	—	76.9%	—	—	26.5%	—	—	—	—	—	—	—	—	—	—	30.7	12.5s	Jan. 2025	Apr. 2026	Google
Gemma 4 26B-A4B	🇺🇸	Open	262.1k	$0.13	$0.40	96c/s	1,229	36.1	32.5	–	–	–	20.7	18.4	18.0	33.1	33.1	32.8	82.3%	—	—	—	86.3%	—	—	—	73.8%	—	—	17.2%	—	—	—	—	—	—	—	—	—	—	25.2	2.5s	Jan. 2025	Apr. 2026	Google
MiniMax M2.7	🇨🇳	Open	204.8k	$0.30	$1.20	66c/s	1,180	50.8	–	38.2	–	–	–	23.5	–	24.1	24.4	–	—	—	—	—	—	—	—	—	—	—	—	—	—	—	46.3%	—	—	—	—	—	—	56.2%	-	3.7s	-	Mar. 2026	MiniMax
DeepSeek-V4-Flash-Max	🇨🇳	Open	1.0M	$0.14	$0.28	59c/s	1,144	50.8	51.6	36.9	23.2	–	31.1	27.0	7.7	34.3	34.3	45.1	88.1%	—	79.0%	—	—	—	73.2%	—	—	—	69.0%	45.1%	34.1%	—	47.8%	—	—	—	—	—	—	52.6%	284	8.5s	-	Apr. 2026	DeepSeek
GLM-4.6	🇨🇳	Open	-	-	-	20c/s	1,144	37.5	34.1	18.8	6.6	–	12.4	–	–	–	–	–	81.0%	93.9%	68.0%	—	—	—	45.1%	—	—	—	—	17.2%	—	—	—	40.5%	—	—	—	—	—	—	357	2.4s	-	Sep. 2025	ZAI
Qwen3.6-27B	🇨🇳	Open	262.1k	$0.60	$3.60	146c/s	1,121	45.1	42.6	33.8	–	–	29.8	23.6	27.8	46.9	47.0	46.0	87.8%	—	77.2%	—	—	82.9%	—	78.4%	75.8%	—	—	24.0%	—	—	—	—	—	—	—	—	—	53.5%	27.8	10.7s	-	Apr. 2026	Qwen
GLM-4.7	🇨🇳	Open	-	-	-	21c/s	1,064	43.7	44.2	22.1	16.6	–	29.9	11.4	–	36.9	36.9	36.6	85.7%	95.7%	73.8%	—	—	—	52.0%	—	—	—	—	42.8%	—	—	—	33.3%	—	—	—	—	—	—	358	45.0s	-	Dec. 2025	ZAI
LongCat-Flash-Chat	🇨🇳	Open	-	-	-	–	1,039	23.1	22.5	14.7	–	16.5	–	13.2	–	34.9	34.9	34.5	73.2%	61.3%	60.4%	—	—	—	—	—	—	—	—	—	—	—	—	39.5%	—	—	—	—	—	—	560	–	-	Aug. 2025	Meituan
MiniMax M2.5	🇨🇳	Open	1M	$0.30	$1.20	335c/s	1,038	50.8	–	37.4	26.1	–	–	–	–	-1.1	–	–	—	—	80.2%	—	—	—	76.3%	—	—	—	—	—	—	—	—	—	—	—	—	—	—	55.4%	230	2.2s	-	Feb. 2026	MiniMax
DeepSeek-V3.2 (Non-thinking)	🇨🇳	Open	131.1k	$0.28	$0.42	177c/s	1,011	–	–	–	–	–	–	–	–	–	–	–	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	685	1.0s	-	Dec. 2025	DeepSeek
MiniMax M3	🇨🇳	Open	1M	$0.60	$2.40	112c/s	993	54.5	–	45.3	32.2	–	33.2	31.1	–	–	–	40.2	—	—	80.5%	—	—	—	83.5%	—	78.1%	—	74.2%	—	—	—	—	—	—	—	—	—	27.7%	59.0%	-	2.6s	-	Jun. 2026	MiniMax
MiniMax M2	🇨🇳	Open	1M	$0.30	$1.20	245c/s	970	33.4	25.5	22.4	4.7	20.1	6.2	14.4	–	30.4	30.4	30.1	78.0%	78.0%	69.4%	—	—	—	44.0%	—	—	—	—	12.5%	—	—	—	46.3%	—	—	—	36.0%	—	—	230	1.3s	-	Oct. 2025	MiniMax
Kimi K2-Thinking-0905	🇨🇳	Open	-	-	-	11c/s	941	45.1	47.2	25.2	19.5	-7.1	36.4	–	–	26.3	26.2	40.7	84.5%	100.0%	71.3%	—	—	—	60.2%	—	—	—	—	51.0%	—	—	—	47.1%	—	—	—	44.8%	—	—	1000	25.6s	-	Sep. 2025	MoonshotAI
Mistral Large 3 (675B Instruct 2512)	🇫🇷	Open	262.1k	$0.50	$1.50	211c/s	882	10.0	19.8	-1.3	–	–	–	–	–	–	–	–	43.9%	—	—	—	85.5%	—	—	—	—	—	—	—	23.8%	—	—	—	—	—	—	—	—	—	675	769ms	-	Dec. 2025	Mistral
MiniMax M2.1	🇨🇳	Open	1M	$0.30	$1.20	599c/s	858	40.9	36.1	26.3	18.7	19.4	17.8	21.1	19.6	50.8	50.8	50.5	81.0%	81.0%	67.0%	—	—	—	62.0%	—	—	—	—	22.0%	—	—	43.5%	47.9%	—	—	—	39.0%	—	—	230	2.7s	-	Dec. 2025	MiniMax
DeepSeek-V3.2-Speciale	🇨🇳	Open	-	-	-	–	832	42.7	45.8	22.7	–	–	24.2	13.3	–	–	–	–	—	96.0%	73.1%	—	—	—	—	—	—	—	—	30.6%	—	—	35.2%	—	—	—	—	—	—	—	685	–	-	Dec. 2025	DeepSeek
MiMo-V2-Flash	🇨🇳	Open	-	-	-	–	793	38.6	38.0	22.0	15.7	–	18.4	8.8	19.7	38.8	38.8	38.5	83.7%	94.1%	73.4%	—	—	—	58.3%	—	—	—	—	22.1%	—	—	—	30.5%	—	—	—	—	—	—	309	–	-	Dec. 2025	Xiaomi
GPT OSS 120B High	🇺🇸	Open	131.1k	$0.10	$0.50	64c/s	792	32.2	29.3	–	–	–	–	3.4	–	26.6	26.6	26.3	80.9%	92.5%	—	—	83.8%	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	116.8	16.9s	-	Aug. 2025	OpenAI
LongCat-Flash-Thinking	🇨🇳	Open	-	-	-	–	791	35.7	33.7	17.9	–	23.6	15.9	20.9	–	37.4	33.0	32.7	81.5%	90.6%	59.4%	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	560	–	-	Sep. 2025	Meituan
Qwen3.5-27B	🇨🇳	Open	262.1k	$0.30	$2.40	85c/s	783	41.9	42.4	20.7	20.1	19.2	28.9	14.6	29.6	46.4	46.5	43.7	85.5%	—	72.4%	—	85.9%	82.3%	61.0%	79.5%	75.0%	70.3%	—	48.5%	—	—	—	—	—	—	—	—	—	—	27	23.9s	-	Feb. 2026	Qwen
Qwen3.5-122B-A10B	🇨🇳	Open	262.1k	$0.40	$3.20	141c/s	780	43.0	44.0	25.4	21.8	20.5	30.2	15.4	31.1	49.5	49.5	47.8	86.6%	—	72.0%	—	86.7%	83.9%	63.8%	77.2%	76.9%	70.4%	—	47.5%	—	—	—	—	—	—	—	—	—	—	122	15.3s	-	Feb. 2026	Qwen
GLM-4.7-Flash	🇨🇳	Open	-	-	-	–	759	30.9	28.7	7.0	3.3	–	7.4	9.5	–	–	–	–	75.2%	91.6%	59.2%	—	—	—	42.8%	—	—	—	—	14.4%	—	—	—	—	—	—	—	—	—	—	30	–	-	Jan. 2026	ZAI
LongCat-Flash-Lite	🇨🇳	Open	256k	$0.10	$0.40	–	756	22.5	19.4	10.7	–	17.5	–	14.0	–	21.5	21.5	21.1	66.8%	63.2%	54.4%	—	—	—	—	—	—	—	—	—	—	—	—	33.8%	—	—	—	—	—	—	68.5	–	-	Feb. 2026	Meituan

Showing 1–30 of 186 models

Open LLM Leaderboard highlights

Independent ranking of open-weight large language models — Llama, Qwen, GLM, DeepSeek, Mistral, Kimi and more — by coding-arena score, GPQA Diamond, throughput, latency, and per-token pricing. Updated continuously from provider APIs and verified benchmarks. See the LLM Stats Score methodology for how rankings are computed.

Best for coding (Arena): Claude Opus 4.6 (21.2 arena score)
Best on GPQA Diamond: Claude Mythos Preview (94.6%)
Best on AIME 2025: Gemini 3 Pro (100.0%)
Best on SWE-Bench Verified: Claude Fable 5 (95.0%)
Highest throughput: MiniMax M2.1 (599 tok/s)
Lowest latency: Qwen3 30B A3B (359.75 s)
Cheapest input: Nemotron 3 Nano (30B A3B) ($0.06 / 1M tok)
Largest context window: Grok 4 Fast (2.0M tokens)

FAQ

Common questions about the open llm leaderboard

What is the best LLM right now?

Based on coding-arena performance — the most discriminating signal at the frontier — Claude Opus 4.6 currently leads. For knowledge-heavy reasoning (GPQA Diamond), Claude Mythos Preview scores highest. Choose by axis rather than a single ranking — see the highlights above for per-metric leaders.

How does the Open LLM Leaderboard rank models?

Models are sorted by coding-arena score (when available), then by GPQA Diamond. Each row aggregates verified benchmark results, provider-reported pricing, and live performance metrics (output throughput and time-to-first-token) sampled across the major API providers. See the LLM Stats Score methodology for the full weighting and refresh cadence.

How many models are tracked?

This leaderboard tracks 316 canonical models across every major lab and provider. New releases typically appear within hours.

Where does pricing data come from?

Per-model input/output pricing is pulled from each provider's public API price list and verified against billing samples from the LLM Stats proxy. When a model is hosted by multiple providers, the cheapest available rate is shown by default.

How is performance measured?

Output throughput (tokens/second) and time-to-first-token are measured by routing standardized prompts through each provider's API and averaging over a 7-day rolling window. Numbers update hourly. Per-model splits live on each model detail page.

How often does the data update?

Pricing and model metadata revalidate every hour. Live performance metrics update on a 7-day rolling average. Benchmark scores update when a new verified result is published or a new evaluation lands on LLM Stats.