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 →


GLM-5	🇨🇳	Open	200k	$1.00	$3.20	–	1,581	52.2	–	37.3	26.5	–	–	26.7	–	–	–	–	—	—	77.8%	—	—	—	75.9%	—	—	—	67.8%	—	—	—	—	—	—	—	—	—	—	—	744	–	-	Feb. 2026	ZAI
GLM-5.1	🇨🇳	Open	200k	$1.40	$4.40	–	1,539	55.1	47.2	45.1	30.1	–	39.5	30.6	–	–	–	–	86.2%	—	—	—	—	—	79.3%	—	—	—	71.8%	52.3%	—	—	40.7%	—	—	—	—	—	—	58.4%	754	–	-	Apr. 2026	ZAI
Kimi K2.5	🇨🇳	Open	262.1k	$0.60	$3.00	71c/s	1,462	50.4	47.9	32.5	30.4	–	35.9	14.6	38.3	47.7	47.7	48.5	87.6%	96.1%	76.8%	—	—	—	74.9%	77.5%	78.5%	—	—	50.2%	—	—	—	—	—	—	—	48.7%	—	50.7%	1000	66.0s	-	Jan. 2026	MoonshotAI
Kimi K2.6	🇨🇳	Open	262.1k	$0.95	$4.00	19c/s	1,254	59.1	50.7	45.6	38.0	–	38.0	33.4	–	–	–	–	90.5%	—	80.2%	—	—	—	86.3%	86.7%	80.1%	—	—	36.4%	—	—	50.0%	—	—	—	—	52.2%	27.9%	58.6%	1000	35.7s	-	Apr. 2026	MoonshotAI
Qwen3.5-397B-A17B	🇨🇳	Open	262.1k	$0.60	$3.60	185c/s	1,207	49.3	46.7	31.0	25.8	28.7	29.0	22.8	38.8	55.4	55.4	54.3	88.4%	—	76.4%	—	88.5%	—	69.0%	—	—	—	—	28.7%	—	—	38.3%	—	—	—	—	—	—	—	397	11.4s	-	Feb. 2026	Qwen
GLM-4.6	🇨🇳	Open	131.1k	$0.55	$2.19	54c/s	1,144	38.3	34.5	20.5	7.4	–	13.5	–	–	–	–	–	81.0%	93.9%	68.0%	—	—	—	45.1%	—	—	—	—	17.2%	—	—	—	40.5%	—	—	—	—	—	—	357	743ms	-	Sep. 2025	ZAI
Gemma 4 26B-A4B	🇺🇸	Open	262.1k	$0.13	$0.40	107c/s	1,141	36.7	33.1	–	–	–	21.1	18.4	15.9	32.6	32.6	31.0	82.3%	—	—	—	86.3%	—	—	—	73.8%	—	—	17.2%	—	—	—	—	—	—	—	—	—	—	25.2	1.5s	Jan. 2025	Apr. 2026	Google
Qwen3 VL 235B A22B Instruct	🇨🇳	Open	262.1k	$0.30	$1.49	106c/s	1,114	26.2	28.7	–	–	19.4	23.7	4.7	23.8	32.5	34.1	32.7	—	74.7%	—	—	—	—	—	62.1%	68.1%	62.0%	—	—	51.9%	66.7%	—	—	—	—	—	—	—	—	236	4.0s	-	Sep. 2025	Qwen
DeepSeek-V4-Pro-Max	🇨🇳	Open	1.0M	$1.74	$3.48	70c/s	1,080	57.8	53.8	45.0	33.5	–	33.5	35.2	20.0	45.2	45.7	48.6	90.1%	—	80.6%	—	—	—	83.4%	—	—	—	73.6%	48.2%	57.9%	—	51.8%	—	—	—	—	—	—	55.4%	1600	95.0s	-	Apr. 2026	DeepSeek
GLM-4.7	🇨🇳	Open	204.8k	$0.60	$2.20	44c/s	1,066	44.3	44.1	23.2	16.8	–	29.7	12.5	–	36.6	36.5	36.3	85.7%	95.7%	73.8%	—	—	—	52.0%	—	—	—	—	42.8%	—	—	—	33.3%	—	—	—	—	—	—	358	8.2s	-	Dec. 2025	ZAI
Gemma 4 31B	🇺🇸	Open	262.1k	$0.14	$0.40	70c/s	979	45.2	39.9	–	–	–	27.8	20.1	29.8	42.4	42.4	36.7	84.3%	—	—	—	88.4%	—	—	—	76.9%	—	—	26.5%	—	—	—	—	—	—	—	—	—	—	30.7	5.9s	Jan. 2025	Apr. 2026	Google
MiniMax M2.5	🇨🇳	Open	1M	$0.30	$1.20	490c/s	955	53.2	–	38.9	27.0	–	–	–	–	-1.6	–	–	—	—	80.2%	—	—	—	76.3%	—	—	—	—	—	—	—	—	—	—	—	—	—	—	55.4%	230	2.3s	-	Feb. 2026	MiniMax
Kimi K2-Thinking-0905	🇨🇳	Open	-	-	-	–	941	45.7	47.3	26.6	19.9	-6.3	36.8	–	–	26.4	26.3	39.4	84.5%	100.0%	71.3%	—	—	—	60.2%	—	—	—	—	51.0%	—	—	—	47.1%	—	—	—	44.8%	—	—	1000	–	-	Sep. 2025	MoonshotAI
MiniMax M2.1	🇨🇳	Open	1M	$0.30	$1.20	639c/s	908	41.6	36.7	27.6	19.1	18.2	18.6	21.1	20.1	51.5	51.5	51.0	81.0%	81.0%	67.0%	—	—	—	62.0%	—	—	—	—	22.0%	—	—	43.5%	47.9%	—	—	—	39.0%	—	—	230	2.0s	-	Dec. 2025	MiniMax
DeepSeek-V4-Flash-Max	🇨🇳	Open	1.0M	$0.14	$0.28	122c/s	842	52.3	51.5	38.8	23.9	–	31.0	28.3	7.9	36.5	36.5	44.7	88.1%	—	79.0%	—	—	—	73.2%	—	—	—	69.0%	45.1%	34.1%	—	47.8%	—	—	—	—	—	—	52.6%	284	18.5s	-	Apr. 2026	DeepSeek
MiniMax M2	🇨🇳	Open	1M	$0.30	$1.20	161c/s	834	34.3	25.9	23.9	5.3	19.1	6.7	14.5	–	29.9	29.9	29.6	78.0%	78.0%	69.4%	—	—	—	44.0%	—	—	—	—	12.5%	—	—	—	46.3%	—	—	—	36.0%	—	—	230	1.1s	-	Oct. 2025	MiniMax
DeepSeek-V3.2-Speciale	🇨🇳	Open	-	-	-	–	832	43.0	46.2	23.5	–	–	25.4	13.1	–	–	–	–	—	96.0%	73.1%	—	—	—	—	—	—	—	—	30.6%	—	—	35.2%	—	—	—	—	—	—	—	685	–	-	Dec. 2025	DeepSeek
MiniMax M2.7	🇨🇳	Open	204.8k	$0.30	$1.20	161c/s	831	53.7	–	40.1	–	–	–	24.3	–	31.2	31.5	–	—	—	—	—	—	—	—	—	—	—	—	—	—	—	46.3%	—	—	—	—	—	—	56.2%	-	5.2s	-	Mar. 2026	MiniMax
MiMo-V2-Flash	🇨🇳	Open	-	-	-	–	793	38.8	38.1	23.6	15.9	–	19.3	9.9	22.1	38.6	38.6	38.3	83.7%	94.1%	73.4%	—	—	—	58.3%	—	—	—	—	22.1%	—	—	—	30.5%	—	—	—	—	—	—	309	–	-	Dec. 2025	Xiaomi
LongCat-Flash-Thinking	🇨🇳	Open	-	-	-	–	791	36.3	34.2	19.4	–	23.8	16.6	21.6	–	36.6	32.5	32.2	81.5%	90.6%	59.4%	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	560	–	-	Sep. 2025	Meituan
LongCat-Flash-Chat	🇨🇳	Open	128k	$0.30	$1.20	171c/s	772	23.9	22.9	16.2	–	17.8	–	14.5	–	35.0	35.0	34.5	73.2%	61.3%	60.4%	—	—	—	—	—	—	—	—	—	—	—	—	39.5%	—	—	—	—	—	—	560	5.9s	-	Aug. 2025	Meituan
GLM-4.7-Flash	🇨🇳	Open	128k	$0.07	$0.40	9c/s	759	32.0	29.3	9.0	3.9	–	8.7	10.2	–	–	–	–	75.2%	91.6%	59.2%	—	—	—	42.8%	—	—	—	—	14.4%	—	—	—	—	—	—	—	—	—	—	30	24.5s	-	Jan. 2026	ZAI
DeepSeek-V3.2-Exp	🇨🇳	Open	-	-	-	–	750	36.2	33.7	22.3	1.9	–	16.5	–	–	39.1	39.1	38.8	79.9%	89.3%	67.8%	—	—	—	40.1%	—	—	—	—	19.8%	97.1%	—	—	37.7%	—	—	—	—	—	—	685	–	-	Sep. 2025	DeepSeek
GLM-4.5	🇨🇳	Open	-	-	-	–	744	34.4	34.5	21.3	-4.2	25.2	8.2	25.8	–	42.0	37.9	37.7	79.1%	—	64.2%	—	—	—	26.4%	—	—	—	—	14.4%	—	—	—	37.5%	79.7%	—	—	41.7%	—	—	355	–	-	Jul. 2025	ZAI
LongCat-Flash-Lite	🇨🇳	Open	256k	$0.10	$0.40	272c/s	712	23.4	20.1	12.3	–	18.7	–	15.4	–	22.3	22.3	21.9	66.8%	63.2%	54.4%	—	—	—	—	—	—	—	—	—	—	—	—	33.8%	—	—	—	—	—	—	68.5	6.6s	-	Feb. 2026	Meituan
DeepSeek-V3.2 (Non-thinking)	🇨🇳	Open	131.1k	$0.28	$0.42	136c/s	705	–	–	–	–	–	–	–	–	–	–	–	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	685	1.2s	-	Dec. 2025	DeepSeek
Qwen3.5-122B-A10B	🇨🇳	Open	262.1k	$0.40	$3.20	84c/s	698	43.5	44.4	26.6	22.2	24.3	31.3	14.1	32.4	49.9	49.9	48.0	86.6%	—	72.0%	—	86.7%	83.9%	63.8%	77.2%	76.9%	70.4%	—	47.5%	—	—	—	—	—	—	—	—	—	—	122	14.5s	-	Feb. 2026	Qwen
Mistral Large 3 (675B Instruct 2512)	🇫🇷	Open	262.1k	$0.50	$1.50	174c/s	644	9.1	19.9	0.1	–	–	–	–	–	–	–	–	43.9%	—	—	—	85.5%	—	—	—	—	—	—	—	23.8%	—	—	—	—	—	—	—	—	—	675	391ms	-	Dec. 2025	Mistral
Step-3.5-Flash	🇨🇳	Open	65.5k	$0.10	$0.40	201c/s	595	49.9	47.1	28.9	22.4	–	–	19.1	–	–	–	–	—	97.3%	74.4%	—	—	—	69.0%	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	196	9.0s	-	Feb. 2026	StepFun
Qwen3-Coder	🇨🇳	Open	-	-	-	–	585	–	–	–	–	–	–	–	–	–	–	–	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	480	–	-	Jan. 2025	Qwen

Showing 1–30 of 179 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): Gemini 3.1 Pro (20.9 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 Mythos Preview (93.9%)
Highest throughput: Mercury 2 (1212 tok/s)
Lowest latency: Mistral Large 3 (675B Instruct 2512) (390.50 s)
Cheapest input: Nemotron 3 Nano (30B A3B) ($0.06 / 1M tok)
Largest context window: Grok-4.20 Beta Non-Reasoning (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 — Gemini 3.1 Pro 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 298 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.