Benchmarks Comparison

Benchmark Comparison

Compare AI model performance across standardized tests

Individual Benchmark Results

This table shows performance on specific standardized tests (MMLU, HumanEval, GPQA, etc.) rather than category aggregations. Scores represent pass rates or accuracy percentages on each benchmark.

Filters & View Options

203 models

Select benchmarks to display (showing first 10):

Model	AIME	AA Coding Index	AAII	AA Math Index	DROP	GPQA	HLE	HumanEval	LiveCodeBench	MATH-500
claude-3-5-sonnet-20240620	15.7	30.2	29.9	—	87.1	59.7	3.9	92.0	38.1	77.1
claude-3-haiku-20240307	1.0	—	9.6	—	78.4	33.3	—	75.9	15.4	39.4
claude-3-opus-20240229	3.3	19.5	20.6	—	83.1	49.6	3.1	84.9	27.9	64.1
claude-3.5-haiku	3.3	—	20.2	—	83.1	41.2	3.5	88.1	31.4	72.1
claude-3.7-sonnet	22.3	32.3	33.5	21.0	—	65.6	4.8	—	39.4	85.0
claude-opus-4	56.3	—	42.3	36.3	—	70.1	5.9	—	54.2	94.1
claude-opus-4.1	—	46.1	59.3	80.3	—	80.9	11.9	—	65.4	—
claude-sonnet-4	77.3	45.1	56.5	74.3	—	77.7	9.6	—	65.5	99.1
claude-sonnet-4.5	—	49.8	62.7	88.0	—	83.4	17.3	—	71.4	—
mistral-7b-instruct-v0.2	0.0	—	1.0	—	—	17.7	4.3	—	4.6	12.1
DeepSeek-R1	89.3	44.1	52.0	76.0	—	81.3	14.9	—	77.0	98.3
DeepSeek-R1-Distill-Llama-70B	67.0	—	30.8	53.7	—	65.2	6.1	—	26.6	93.5
DeepSeek-R1-Distill-Qwen-32B	68.7	—	32.7	63.0	—	61.8	5.5	—	27.0	94.1
DeepSeek-V3	25.3	25.9	32.5	26.0	91.6	57.4	3.6	—	35.9	88.7
deepseek-chat-v3-0324	—	39.0	44.8	49.7	—	73.5	6.3	—	57.7	—
deepseek-chat-v3.1	—	39.0	44.8	49.7	—	73.5	6.3	—	57.7	—
deepseek-r1-0528	—	—	—	—	—	81.0	—	—	—	—
devstral-small	0.3	18.5	27.2	29.3	—	41.4	3.7	—	25.4	63.5
devstral-small-2505	6.7	—	19.6	—	—	43.4	4.0	—	25.8	68.4
gemma-3-12b-it	22.0	10.6	20.4	18.3	—	40.9	4.8	85.4	13.7	85.3
gemma-3-27b-it	25.3	12.8	21.6	20.7	—	42.6	4.7	87.8	13.7	88.3
gemma-3-4b-it	6.3	6.4	14.7	12.7	—	29.9	5.2	71.3	11.2	76.6
gpt-oss-120b	—	41.2	58.0	93.4	—	79.2	18.5	—	65.3	—
gpt-oss-20b	—	32.8	43.3	61.7	—	71.5	8.5	—	57.2	—
kimi-k2	—	38.1	50.4	57.3	—	76.3	6.3	94.5	61.0	—
llama-3.1-405b-instruct	21.3	22.2	25.7	3.0	84.8	51.1	4.2	89.0	30.5	70.3
llama-3.1-70b-instruct	17.3	17.6	22.6	4.0	79.6	41.3	4.6	80.5	23.2	64.9
llama-3.1-8b-instruct	7.7	8.5	16.9	4.3	59.5	28.1	5.1	72.6	11.6	51.9
llama-3.1-nemotron-70b-instruct	24.7	14.8	23.6	11.0	—	46.5	4.6	—	16.9	73.3
llama-3.2-3b-instruct	6.7	—	11.2	3.3	—	32.8	5.2	—	8.3	48.9
llama-3.3-70b-instruct	30.0	19.2	27.9	7.7	—	50.2	4.0	88.4	28.8	77.3
llama-4-maverick	39.0	26.4	35.8	19.3	—	68.5	4.8	—	39.7	88.9
llama-4-scout	28.3	16.1	28.1	14.0	—	57.9	4.3	—	29.9	84.4
mistral-7b-instruct	0.0	—	1.0	—	—	17.7	4.3	—	4.6	12.1
mistral-nemo	0.3	—	5.2	—	—	31.4	4.4	—	5.7	39.5
mistral-small-24b-instruct-2501	—	—	—	—	—	45.3	—	84.8	—	—
mistral-small-3.1-24b-instruct	6.3	—	13.0	—	—	38.1	4.3	—	14.1	56.3
mistral-small-3.2-24b-instruct	6.3	—	13.0	—	—	38.1	4.3	—	14.1	56.3
mixtral-8x7b-instruct	0.0	—	2.6	—	—	29.2	4.5	—	6.6	29.9
nemotron-nano-9b-v2	—	30.6	36.1	62.3	—	55.7	4.0	—	70.1	—
phi-4	14.3	17.6	24.6	18.0	75.5	56.8	4.1	82.6	23.1	81.0
phi-4-multimodal-instruct	9.3	—	12.4	—	—	31.5	4.4	—	13.1	69.3
phi-4-reasoning-plus	—	—	—	—	—	68.9	—	—	—	—
qwen-2.5-72b-instruct	16.0	19.5	29.0	14.0	—	49.0	4.2	86.6	27.6	85.8
qwen-2.5-7b-instruct	—	—	—	—	—	36.4	—	84.8	—	—
qwen3-14b	28.0	19.8	29.2	58.0	—	47.0	4.2	—	28.0	87.1
qwen3-235b-a22b	32.7	23.3	29.9	23.7	—	61.3	4.7	—	34.3	90.2
qwen3-235b-a22b-2507	32.7	23.3	29.9	23.7	—	61.3	4.7	—	34.3	90.2
qwen3-235b-a22b-thinking-2507	—	—	—	—	—	81.1	—	—	—	—
qwen3-30b-a3b	72.7	29.2	37.0	66.3	—	65.9	6.8	—	51.5	97.5
qwen3-32b	80.7	30.9	38.7	73.0	—	66.8	8.3	—	54.6	96.1
qwen3-coder	47.7	37.4	42.3	39.3	—	61.8	4.4	—	58.5	94.2
qwq-32b	78.0	—	37.9	29.0	—	65.2	8.2	—	63.1	95.7
deepseek-chat	—	39.0	44.8	49.7	—	73.5	6.3	—	57.7	—
deepseek-reasoner	—	47.2	54.0	89.7	—	77.9	13.0	—	78.4	—
deepseek-r1	89.3	44.1	52.0	76.0	—	81.3	14.9	—	77.0	98.3
deepseek-r1-05-28	89.3	44.1	52.0	76.0	—	81.3	14.9	—	77.0	98.3
deepseek-v3	25.3	25.9	32.5	26.0	91.6	57.4	3.6	—	35.9	88.7
deepseek-v3-03-24	25.3	25.9	32.5	26.0	91.6	57.4	3.6	—	35.9	88.7
gpt-oss-120b	—	41.2	58.0	93.4	—	79.2	18.5	—	65.3	—
gpt-oss-20b	—	32.8	43.3	61.7	—	71.5	8.5	—	57.2	—
kimi-k2-instruct	—	—	—	—	—	75.1	—	93.3	—	—
qwen3-235b-a22b	32.7	23.3	29.9	23.7	—	61.3	4.7	—	34.3	90.2
qwen3-coder-480b-a35b-instruct	47.7	37.4	42.3	39.3	—	61.8	4.4	—	58.5	94.2
gemini-2.0-flash	33.0	23.4	33.6	21.7	—	62.2	5.3	—	33.4	93.0
gemini-2.0-flash-lite	—	—	—	—	—	51.5	—	—	—	—
gemini-2.5-flash	50.0	30.0	40.4	60.3	—	82.8	5.1	—	49.5	93.2
gemini-2.5-pro	88.7	49.3	59.6	87.7	—	83.7	21.1	—	80.1	96.7
gemini-2.5-pro-preview	—	—	—	—	—	86.4	—	—	—	—
deepseek-r1-distill-llama-70b	67.0	—	30.8	53.7	—	65.2	6.1	—	26.6	93.5
gemma-2-9b-it	0.0	—	7.8	—	—	31.1	3.9	40.2	12.6	51.7
gpt-oss-120b	—	41.2	58.0	93.4	—	79.2	18.5	—	65.3	—
gpt-oss-20b	—	32.8	43.3	61.7	—	71.5	8.5	—	57.2	—
kimi-k2	—	38.1	50.4	57.3	—	76.3	6.3	94.5	61.0	—
kimi-k2-0905	—	38.1	50.4	57.3	—	76.3	6.3	94.5	61.0	—
llama-3.1-8b-instruct	7.7	8.5	16.9	4.3	59.5	28.1	5.1	72.6	11.6	51.9
llama-3.3-70b-instruct	30.0	19.2	27.9	7.7	—	50.2	4.0	88.4	28.8	77.3
llama-4-maverick	39.0	26.4	35.8	19.3	—	68.5	4.8	—	39.7	88.9
llama-4-scout	28.3	16.1	28.1	14.0	—	57.9	4.3	—	29.9	84.4
qwen3-32b	80.7	30.9	38.7	73.0	—	66.8	8.3	—	54.6	96.1
codestral-2501	4.3	—	13.2	—	—	31.2	4.5	—	24.3	60.7
codestral-2508	4.3	—	13.2	—	—	31.2	4.5	—	24.3	60.7
devstral-medium	6.7	23.9	27.9	4.7	—	49.2	3.8	—	33.7	70.7
ministral-3b	0.0	5.4	5.2	—	—	26.0	5.5	—	6.9	53.7
mistral-large	0.0	—	11.9	—	—	35.1	3.4	—	17.8	52.7
mistral-large-2.1	0.0	—	11.9	—	—	35.1	3.4	—	17.8	52.7
mistral-large-2407	9.3	—	22.3	0.0	—	47.2	3.2	—	26.7	71.4
mistral-medium-3	44.0	25.6	34.7	30.3	—	57.8	4.3	—	40.0	90.7
mistral-medium-3.1	3.7	—	8.4	—	—	34.9	3.4	—	9.9	40.5
mistral-nemo	0.3	—	5.2	—	—	31.4	4.4	—	5.7	39.5
mistral-saba	13.0	—	19.6	—	—	42.4	4.1	—	—	67.7
mistral-small	6.3	—	13.0	—	—	38.1	4.3	—	14.1	56.3
mistral-small-3	6.3	—	13.0	—	—	38.1	4.3	—	14.1	56.3
mistral-small-3.1	6.3	—	13.0	—	—	38.1	4.3	—	14.1	56.3
mistral-small-3.2	6.3	—	13.0	—	—	38.1	4.3	—	14.1	56.3
chatgpt-4o-latest	32.7	—	35.6	25.7	—	65.5	5.0	—	42.5	89.3
gpt-3.5-turbo	—	10.7	8.3	—	70.2	30.3	—	68.0	—	44.1
gpt-4.1	43.7	32.2	43.4	34.7	—	66.5	4.6	—	45.7	91.3
gpt-4.1-mini	—	—	—	—	—	65.0	—	—	—	—
gpt-4.1-nano	—	—	—	—	—	50.3	—	—	—	—
gpt-4o	15.0	24.0	27.0	6.0	—	54.3	3.3	—	30.9	75.9
gpt-4o-2024-05-13	15.0	24.0	27.0	6.0	83.4	54.0	3.3	90.2	30.9	75.9
gpt-4o-mini	11.7	—	21.2	14.7	—	42.6	4.0	—	23.4	78.9
gpt-4o-mini-search-preview	—	—	—	—	79.7	40.2	—	87.2	—	—
gpt-5	95.7	52.7	68.5	94.3	—	85.4	26.5	—	84.6	99.4
gpt-5-chat	—	34.7	41.8	48.3	—	68.6	5.8	—	54.3	—
gpt-5-mini	—	51.4	64.3	90.7	—	82.8	19.7	—	83.8	—
gpt-5-nano	—	42.3	51.0	83.7	—	67.6	8.2	—	78.9	—
o1	72.3	38.6	47.2	—	—	76.4	7.7	88.1	67.9	97.0
o1-mini	60.3	—	39.2	—	—	60.1	4.9	92.4	57.6	94.4
o3	90.3	52.2	65.5	88.3	—	83.0	20.0	—	80.8	99.2
o3-mini	77.0	39.4	48.1	—	—	76.0	8.7	—	71.7	97.3
o4-mini	94.0	48.9	59.6	90.7	—	79.9	17.5	—	85.9	98.9
codestral-2501	4.3	—	13.2	—	—	31.2	4.5	—	24.3	60.7
codestral-2508	4.3	—	13.2	—	—	31.2	4.5	—	24.3	60.7
command	0.7	—	5.5	—	—	32.3	4.5	—	12.2	27.9
command-a	9.7	—	28.1	13.0	—	52.7	4.6	—	28.7	81.9
command-r	0.3	—	1.0	—	—	28.9	5.1	—	4.4	14.9
command-r-03-2024	0.7	—	1.0	—	—	28.4	4.8	—	4.8	16.4
command-r-plus	0.0	—	7.1	—	—	33.7	5.0	—	11.1	40.2
command-r-plus-04-2024	0.7	—	5.5	—	—	32.3	4.5	—	12.2	27.9
deephermes-3-mistral-24b-preview	4.7	—	15.5	—	—	38.2	3.9	—	19.5	59.5
deepseek-r1-distill-llama-70b	67.0	—	30.8	53.7	—	65.2	6.1	—	26.6	93.5
deepseek-r1-distill-llama-8b	33.3	—	19.5	41.3	—	49.0	4.2	—	23.3	85.3
deepseek-r1-distill-qwen-14b	66.7	—	29.7	55.7	—	59.1	4.4	—	37.6	94.9
deepseek-r1-distill-qwen-32b	68.7	—	32.7	63.0	—	61.8	5.5	—	27.0	94.1
deepseek-v3.1-base	25.3	25.9	32.5	26.0	91.6	57.4	3.6	—	35.9	88.7
devstral-medium	6.7	23.9	27.9	4.7	—	49.2	3.8	—	33.7	70.7
dolphin3.0-mistral-24b	—	—	—	—	—	44.2	—	—	—	—
dolphin3.0-r1-mistral-24b	—	—	—	—	—	44.2	—	—	—	—
gemini-2.5-flash-lite	50.0	19.9	30.1	35.3	—	64.6	3.7	—	40.0	92.6
gemma-2-9b-it	0.0	—	7.8	—	—	31.1	3.9	40.2	12.6	51.7
gemma-3n-e4b-it	—	—	—	—	—	23.7	—	75.0	—	—
grok-2-1212	13.3	—	24.7	—	—	51.0	3.8	—	26.7	77.8
grok-3-mini-beta	—	—	—	—	—	84.0	—	—	—	—
kimi-k2-0905	—	38.1	50.4	57.3	—	76.3	6.3	94.5	61.0	—
llama-3.1-405b	21.3	22.2	25.7	3.0	84.8	51.1	4.2	89.0	30.5	70.3
llama-3.1-nemotron-ultra-253b-v1	74.7	33.7	38.5	63.7	—	74.4	8.1	—	64.1	95.2
magistral-small-2506	—	—	—	—	—	68.2	—	—	—	—
minimax-m1	81.3	—	41.6	13.7	—	68.2	7.5	—	65.7	97.2
ministral-3b	0.0	5.4	5.2	—	—	26.0	5.5	—	6.9	53.7
mistral-7b-instruct-v0.1	0.0	—	1.0	—	—	17.7	4.3	—	4.6	12.1
mistral-7b-instruct-v0.3	0.0	—	1.0	—	—	17.7	4.3	—	4.6	12.1
mistral-large	0.0	—	11.9	—	—	35.1	3.4	—	17.8	52.7
mistral-large-2407	9.3	—	22.3	0.0	—	47.2	3.2	—	26.7	71.4
mistral-large-2411	0.0	—	11.9	—	—	35.1	3.4	—	17.8	52.7
mistral-medium-3	44.0	25.6	34.7	30.3	—	57.8	4.3	—	40.0	90.7
mistral-medium-3.1	3.7	—	8.4	—	—	34.9	3.4	—	9.9	40.5
mistral-saba	13.0	—	19.6	—	—	42.4	4.1	—	—	67.7
mistral-small	6.3	—	13.0	—	—	38.1	4.3	—	14.1	56.3
mistral-tiny	6.3	—	13.0	—	—	38.1	4.3	—	14.1	56.3
nova-lite-v1	10.7	—	21.4	7.0	80.2	42.6	4.6	85.4	16.7	76.5
nova-micro-v1	8.0	—	17.3	6.0	79.3	37.9	4.7	81.1	14.0	70.3
nova-pro-v1	10.7	—	25.5	7.0	85.4	48.4	3.4	89.0	23.3	78.6
qwen-2.5-coder-32b-instruct	12.0	—	21.8	—	—	41.7	3.8	92.7	29.5	76.7
qwen-turbo	12.0	—	19.1	—	—	41.0	4.2	—	16.3	80.5
qwen3-30b-a3b-instruct-2507	72.7	29.2	37.0	66.3	—	65.9	6.8	—	51.5	97.5
qwen3-30b-a3b-thinking-2507	72.7	29.2	37.0	66.3	—	65.9	6.8	—	51.5	97.5
qwen3-8b	74.7	21.8	28.3	19.0	—	58.9	4.2	—	40.6	90.4
qwen3-max	—	44.7	55.1	80.7	—	76.4	11.1	—	76.7	—
qwq-32b-preview	45.3	—	28.0	—	—	65.2	4.8	—	33.7	91.0
r1-1776	—	—	19.1	—	—	—	—	—	—	95.4
sonar	48.7	—	28.8	—	—	47.1	7.3	—	29.5	81.7
sonar-pro	29.0	—	28.2	—	—	57.8	7.9	—	27.5	74.5
sonar-reasoning	77.0	—	34.2	—	—	62.3	—	—	—	92.1
sonar-reasoning-pro	79.0	—	46.3	—	—	—	—	—	—	95.7
deepseek-chat-v3-0324	—	39.0	44.8	49.7	—	73.5	6.3	—	57.7	—
deepseek-chat-v3.1	—	39.0	44.8	49.7	—	73.5	6.3	—	57.7	—
deepseek-r1-0528	—	—	—	—	—	81.0	—	—	—	—
deepseek-r1-0528-qwen3-8b	89.3	44.1	52.0	76.0	—	81.3	14.9	—	77.0	98.3
gemma-3-27b-it	25.3	12.8	21.6	20.7	—	42.6	4.7	87.8	13.7	88.3
glm-4.5	87.3	43.3	51.3	73.7	—	78.6	12.2	—	73.8	97.9
glm-4.5v	—	20.1	26.0	15.3	—	57.3	3.6	—	35.2	—
gpt-oss-120b	—	41.2	58.0	93.4	—	79.2	18.5	—	65.3	—
kimi-k2	—	38.1	50.4	57.3	—	76.3	6.3	94.5	61.0	—
llama-3.3-70b-instruct	30.0	19.2	27.9	7.7	—	50.2	4.0	88.4	28.8	77.3
llama-4-maverick	39.0	26.4	35.8	19.3	—	68.5	4.8	—	39.7	88.9
mistral-nemo	0.3	—	5.2	—	—	31.4	4.4	—	5.7	39.5
mistral-small-3.2-24b-instruct	6.3	—	13.0	—	—	38.1	4.3	—	14.1	56.3
qwen3-14b	28.0	19.8	29.2	58.0	—	47.0	4.2	—	28.0	87.1
qwen3-235b-a22b-2507	32.7	23.3	29.9	23.7	—	61.3	4.7	—	34.3	90.2
qwen3-235b-a22b-thinking-2507	—	—	—	—	—	81.1	—	—	—	—
qwen3-30b-a3b	72.7	29.2	37.0	66.3	—	65.9	6.8	—	51.5	97.5
qwen3-32b	80.7	30.9	38.7	73.0	—	66.8	8.3	—	54.6	96.1
qwen3-coder	47.7	37.4	42.3	39.3	—	61.8	4.4	—	58.5	94.2
DeepSeek-R1-Distill-Llama-70B	67.0	—	30.8	53.7	—	65.2	6.1	—	26.6	93.5
DeepSeek-R1-Distill-Qwen-1.5B	17.7	—	8.6	22.0	—	33.8	3.3	—	7.0	68.7
DeepSeek-R1-Distill-Qwen-14B	66.7	—	29.7	55.7	—	59.1	4.4	—	37.6	94.9
DeepSeek-V3	25.3	25.9	32.5	26.0	91.6	57.4	3.6	—	35.9	88.7
QwQ-32B-Preview	45.3	—	28.0	—	—	65.2	4.8	—	33.7	91.0
Qwen2-72B-Instruct	14.7	—	18.1	—	—	42.4	3.7	86.0	15.9	70.1
gemma-2-27b-it	29.7	—	17.2	—	—	35.7	3.7	51.8	27.9	54.1
grok-2	—	—	—	—	—	56.0	—	88.4	—	—
grok-3	33.0	26.9	36.0	—	—	84.6	5.1	—	42.5	87.0
grok-3-mini	—	—	—	—	—	84.0	—	—	—	—
grok-4	94.3	55.1	65.3	92.7	—	87.6	23.9	—	81.9	99.0
grok-code-fast-1	—	39.4	48.6	43.3	—	72.7	7.5	—	65.7	—
glm-4.5	87.3	43.3	51.3	73.7	—	78.6	12.2	—	73.8	97.9
glm-4.5-air	67.3	39.4	48.8	80.7	—	74.2	6.8	—	68.4	96.5
glm-4.5-airx	67.3	39.4	48.8	80.7	—	74.2	6.8	—	68.4	96.5
glm-4.5-x	87.3	43.3	51.3	73.7	—	78.6	12.2	—	73.8	97.9
glm-4.5v	—	20.1	26.0	15.3	—	57.3	3.6	—	35.2	—
glm-4.6	—	38.7	44.7	44.3	—	81.0	5.2	—	56.1	—

Showing 203 models with benchmark data. Click column headers to sort. Select benchmarks above to customize view.