oobabooga benchmark

The list is sorted by size (on disk) for each score.

Highlighted = Pareto frontier.

Score	Model	Parameters	Size (GB)	Loader	Additional info
38/48	Meta-Llama-3.1-70B-Instruct-Q4_K_M	70B	42.5	llamacpp_HF	[link]
38/48	Mistral-Large-Instruct-2407.i1-Q4_K_S	123B	69.5	llamacpp_HF	[link]
37/48	Meta-Llama-3.1-70B-Instruct-Q3_K_M	70B	34.3	llamacpp_HF	[link]
37/48	Meta-Llama-3.1-70B-Instruct-Q3_K_L	70B	37.1	llamacpp_HF	[link]
37/48	Meta-Llama-3.1-70B-Instruct-Q3_K_XL	70B	38.1	llamacpp_HF	[link]
37/48	Meta-Llama-3.1-70B-Instruct-Q4_K_L	70B	43.3	llamacpp_HF	[link]
37/48	Qwen2.5-72B-Instruct-Q4_K_M	72B	47.42	llamacpp_HF
37/48	Mistral-Large-Instruct-2407.i1-IQ4_XS	123B	65.4	llamacpp_HF	[link]
37/48	Mistral-Large-Instruct-2407.i1-Q6_K	123B	100.7	llamacpp_HF	[link]
36/48	Meta-Llama-3.1-70B-Instruct-Q3_K_S	70B	30.9	llamacpp_HF	[link]
36/48	Meta-Llama-3.1-70B-Instruct-Q4_K_S	70B	40.3	llamacpp_HF	[link]
36/48	turboderp_Llama-3.1-70B-Instruct-exl2_6.0bpw	70B	54.12	ExLlamav2_HF
35/48	Meta-Llama-3.1-70B-Instruct-IQ3_XS	70B	29.3	llamacpp_HF	[link]
35/48	Meta-Llama-3.1-70B-Instruct-IQ4_XS	70B	37.9	llamacpp_HF	[link]
35/48	hugging-quants_Meta-Llama-3.1-70B-Instruct-AWQ-INT4	70B	39.77	Transformers
35/48	turboderp_Llama-3.1-70B-Instruct-exl2_4.5bpw	70B	41.41	ExLlamav2_HF
35/48	Meta-Llama-3.1-70B-Instruct-Q5_K_S	70B	48.7	llamacpp_HF	[link]
35/48	RYS-XLarge-Q4_K_M	72B	50.7	llamacpp_HF	[link]
35/48	Meta-Llama-3.1-70B-Instruct-Q6_K	70B	57.9	llamacpp_HF	[link]
35/48	Meta-Llama-3.1-70B-Instruct-Q6_K_L	70B	58.4	llamacpp_HF	[link]
35/48	Mistral-Large-Instruct-2411-IQ4_XS	123B	65.4	llamacpp_HF	[link]
35/48	turboderp_Mistral-Large-Instruct-2407-123B-exl2_4.25bpw	123B	65.83	ExLlamav2_HF
34/48	Meta-Llama-3.1-70B-Instruct-Q2_K	70B	26.4	llamacpp_HF	[link]
34/48	Meta-Llama-3.1-70B-Instruct-Q2_K_L	70B	27.4	llamacpp_HF	[link]
34/48	Meta-Llama-3.1-70B-Instruct-IQ3_M	70B	31.9	llamacpp_HF	[link]
34/48	Qwen2.5-32B-Instruct-Q8_0	32B	34.82	llamacpp_HF
34/48	platypus-yi-34b.Q8_0	34B	36.5	llamacpp_HF
34/48	Meta-Llama-3-70B-Instruct-Q4_K_S	70B	40.3	llamacpp_HF	[link]
34/48	Llama-3.1-Nemotron-70B-Instruct-HF-Q4_K_M	70B	42.52	llamacpp_HF
34/48	Meta-Llama-3.1-70B-Instruct-Q5_K_M	70B	50.0	llamacpp_HF	[link]
34/48	Meta-Llama-3.1-70B-Instruct-Q5_K_L	70B	50.6	llamacpp_HF	[link]
34/48	LoneStriker_OpenBioLLM-Llama3-70B-6.0bpw-h6-exl2	70B	54.27	ExLlamav2_HF
33/48	gemma-2-27b-it-Q4_K_S	27B	15.74	llamacpp_HF	[link]
33/48	gemma-2-27b-it-Q4_K_M	27B	16.65	llamacpp_HF	[link]
33/48	gemma-2-27b-it-Q5_K_S	27B	18.88	llamacpp_HF	[link]
33/48	gemma-2-27b-it-Q5_K_M	27B	19.41	llamacpp_HF	[link]
33/48	microsoft_Phi-3.5-MoE-instruct	16x3.8B	20.94	Transformers	--load-in-4bit
33/48	ISTA-DASLab_Meta-Llama-3-70B-Instruct-AQLM-2Bit-1x16	70B	21.92	Transformers
33/48	gemma-2-27b-it-Q6_K	27B	22.34	llamacpp_HF	[link]
33/48	gemma-2-27b-it-Q6_K_L	27B	22.63	llamacpp_HF	[link]
33/48	Meta-Llama-3-70B-Instruct-IQ3_XXS	70B	27.5	llamacpp_HF	[link]
33/48	gemma-2-27b-it-Q8_0	27B	28.94	llamacpp_HF	[link]
33/48	Meta-Llama-3-70B-Instruct-IQ3_XS	70B	29.3	llamacpp_HF	[link]
33/48	Meta-Llama-3-70B-Instruct-Q3_K_S	70B	30.9	llamacpp_HF	[link]
33/48	01-ai_Yi-1.5-34B-Chat	34B	34.39	Transformers	--load-in-8bit
33/48	01-ai_Yi-1.5-34B-Chat-16K	34B	34.39	Transformers	--load-in-8bit
33/48	Undi95_Meta-Llama-3-70B-Instruct-hf	70B	35.28	Transformers	--load-in-4bit
33/48	Mistral-Large-Instruct-2407-IQ2_XS	123B	36.1	llamacpp_HF	[link]
33/48	dolphin-2.9.3-Yi-1.5-34B-32k-Q8_0	34B	36.5	llamacpp_HF	[link]
33/48	Llama3-TenyxChat-70B.i1-Q4_K_S	70B	40.3	llamacpp_HF
33/48	Smaug-Llama-3-70B-Instruct.i1-Q4_K_S	70B	40.3	llamacpp_HF
33/48	Meta-Llama-3-70B-Instruct.Q4_K_M	70B	42.5	llamacpp_HF
33/48	turboderp_Cat-Llama-3-70B-instruct-exl2_5.0bpw	70B	45.7	ExLlamav2_HF
33/48	turboderp_Llama-3-70B-Instruct-exl2_5.0bpw	70B	45.72	ExLlamav2_HF
33/48	magnum-72b-v1-Q4_K_M	72B	47.4	llamacpp_HF	[link]
33/48	turboderp_Llama-3-70B-Instruct-exl2_6.0bpw	70B	54.17	ExLlamav2_HF
33/48	LoneStriker_dolphin-2.9-llama3-70b-6.0bpw-h6-exl2	70B	54.27	ExLlamav2_HF
33/48	google_gemma-2-27b-it	27B	54.45	Transformers	--bf16 --use_eager_attention
33/48	Meta-Llama-3-70B-Instruct.Q8_0	70B	75.0	llamacpp_HF
33/48	Meta-Llama-3.1-405B-Instruct.i1-IQ2_XXS	405B	109.0	llamacpp_HF	[link]
32/48	Phi-3-medium-128k-instruct-Q4_K_S	14B	7.95	llamacpp_HF	[link]
32/48	Phi-3-medium-128k-instruct-Q5_K_S	14B	9.62	llamacpp_HF	[link]
32/48	Phi-3-medium-128k-instruct-Q6_K	14B	11.45	llamacpp_HF	[link]
32/48	gemma-2-27b-it-IQ4_XS	27B	14.81	llamacpp_HF	[link]
32/48	gemma-2-27b-it-Q3_K_XL	27B	14.81	llamacpp_HF	[link]
32/48	Phi-3-medium-128k-instruct-Q8_0	14B	14.83	llamacpp_HF	[link]
32/48	gemma-2-27b-it-Q4_K_L	27B	16.93	llamacpp_HF	[link]
32/48	Meta-Llama-3-70B-Instruct-IQ2_S	70B	22.2	llamacpp_HF	[link]
32/48	Meta-Llama-3-70B-Instruct-IQ2_M	70B	24.1	llamacpp_HF	[link]
32/48	turboderp_gemma-2-27b-it-exl2_8.0bpw	27B	26.7	ExLlamav2_HF	--no_flash_attn --no_xformers --no_sdpa
32/48	Meta-Llama-3-70B-Instruct-IQ3_S	70B	30.9	llamacpp_HF	[link]
32/48	Meta-Llama-3-70B-Instruct-Q3_K_M	70B	34.3	llamacpp_HF	[link]
32/48	Meta-Llama-3-70B-Instruct-Q3_K_L	70B	37.1	llamacpp_HF	[link]
32/48	turboderp_Llama-3.1-70B-Instruct-exl2_4.0bpw	70B	37.15	ExLlamav2_HF
32/48	Meta-Llama-3-70B-Instruct-IQ4_XS	70B	37.9	llamacpp_HF	[link]
32/48	Meta-Llama-3-70B-Instruct-IQ4_NL	70B	40.1	llamacpp_HF	[link]
32/48	Llama-3-Giraffe-70B-Instruct.i1-Q4_K_S	70B	40.3	llamacpp_HF	[link]
32/48	Athene-70B.i1-Q4_K_M	70B	42.5	llamacpp_HF	[link]
32/48	Qwen2-72B-Instruct-Q4_K_M	72B	47.4	llamacpp_HF	[link]
32/48	dnhkng_RYS-Gemma-2-27b-it	27B	58.98	Transformers
31/48	Phi-3-medium-4k-instruct-IQ3_XXS	14B	5.45	llamacpp_HF	[link]
31/48	Phi-3-medium-128k-instruct-IQ3_M	14B	6.47	llamacpp_HF	[link]
31/48	Phi-3-medium-128k-instruct-IQ4_XS	14B	7.47	llamacpp_HF	[link]
31/48	Phi-3-medium-128k-instruct-IQ4_NL	14B	7.90	llamacpp_HF	[link]
31/48	Phi-3-medium-128k-instruct-Q4_K_M	14B	8.57	llamacpp_HF	[link]
31/48	Phi-3-medium-128k-instruct-Q5_K_M	14B	10.07	llamacpp_HF	[link]
31/48	gemma-2-27b-it-Q3_K_S	27B	12.17	llamacpp_HF	[link]
31/48	gemma-2-27b-it-IQ3_M	27B	12.45	llamacpp_HF	[link]
31/48	gemma-2-27b-it-Q3_K_L	27B	14.52	llamacpp_HF	[link]
31/48	gemma-2-27b-it-Q5_K_L	27B	19.69	llamacpp_HF	[link]
31/48	Meta-Llama-3-70B-Instruct-IQ2_XS	70B	21.1	llamacpp_HF	[link]
31/48	Meta-Llama-3-70B-Instruct-IQ2_XS	70B	21.1	llamacpp_HF	[link]
31/48	Meta-Llama-3-70B-Instruct-Q2_K	70B	26.4	llamacpp_HF	[link]
31/48	microsoft_Phi-3-medium-128k-instruct	14B	27.92	Transformers
31/48	Meta-Llama-3-70B-Instruct-IQ3_M	70B	31.9	llamacpp_HF	[link]
31/48	miqu-1-70b.q4_k_m	70B	41.4	llamacpp_HF
31/48	Reflection-Llama-3.1-70B-Q4_K_M	70B	42.52	llamacpp_HF
31/48	miqu-1-70b.q5_K_M	70B	48.8	llamacpp_HF
31/48	oobabooga_miqu-1-70b-sf-EXL2-6.000b	70B	52.1	ExLlamav2_HF
31/48	cloudyu_Phoenix_DPO_60B	60B	60.81	Transformers	--load-in-8bit
31/48	Meta-Llama-3-120B-Instruct.Q4_K_M	120B	73.22	llamacpp_HF	[link]
30/48	Phi-3-medium-128k-instruct-IQ3_XXS	14B	5.45	llamacpp_HF	[link]
30/48	Phi-3-medium-128k-instruct-IQ3_S	14B	6.06	llamacpp_HF	[link]
30/48	Phi-3-medium-128k-instruct-Q3_K_L	14B	7.49	llamacpp_HF	[link]
30/48	gemma-2-27b-it-Q3_K_M	27B	13.42	llamacpp_HF	[link]
30/48	google_gemma-2-27b-it	27B	13.61	Transformers	--load-in-4bit --bf16 --use_eager_attention
30/48	Qwen_Qwen2.5-7B-Instruct	7B	15.23	Transformers
30/48	Meta-Llama-3.1-70B-Instruct-IQ2_M	70B	24.1	llamacpp_HF	[link]
30/48	Meta-Llama-3-70B-Instruct-abliterated-v3.5.i1-Q4_K_S	70B	40.3	llamacpp_HF	[link]
30/48	Senku-70B-Full-Q4_K_M	70B	41.4	llamacpp_HF
30/48	microsoft_Phi-3.5-MoE-instruct	16x3.8B	41.87	Transformers	--load-in-8bit
30/48	Llama3-70B-ShiningValiant2.i1-Q4_K_M	70B	42.5	llamacpp_HF	[link]
30/48	Rhea-72b-v0.5-Q4_K_M	72B	43.8	llamacpp_HF
30/48	qwen1_5-72b-chat-q4_k_m	72B	44.2	llamacpp_HF
30/48	Phi-3-medium-128k-instruct-f32	14B	55.80	llamacpp_HF	[link]
30/48	Dracones_WizardLM-2-8x22B_exl2_4.0bpw	8x22B	70.67	ExLlamav2_HF
30/48	turboderp_Mixtral-8x22B-Instruct-v0.1-exl2_4.0bpw	8x22B	70.68	ExLlamav2_HF
30/48	miquliz-120b-v2.0.Q4_K_M	120B	72.1	llamacpp_HF
30/48	falcon-180b-chat.Q4_K_M	180B	108.6	llamacpp_HF
29/48	Phi-3-medium-128k-instruct-IQ3_XS	14B	5.81	llamacpp_HF	[link]
29/48	gemma-2-9b-it-Q5_K_S	9B	6.48	llamacpp_HF	[link]
29/48	Phi-3-medium-128k-instruct-Q3_K_M	14B	6.92	llamacpp_HF	[link]
29/48	Phi-3-medium-4k-instruct-Q4_K_S	14B	7.95	llamacpp_HF	[link]
29/48	Phi-3-medium-4k-instruct-Q4_K_M	14B	8.57	llamacpp_HF	[link]
29/48	Phi-3-medium-4k-instruct-Q5_K_S	14B	9.62	llamacpp_HF	[link]
29/48	Phi-3-medium-4k-instruct-Q5_K_M	14B	10.07	llamacpp_HF	[link]
29/48	Phi-3-medium-4k-instruct-Q6_K	14B	11.45	llamacpp_HF	[link]
29/48	gemma-2-27b-it-IQ3_XS	27B	11.55	llamacpp_HF	[link]
29/48	Phi-3-medium-4k-instruct-Q8_0	14B	14.83	llamacpp_HF	[link]
29/48	google_gemma-2-9b-it	9B	18.48	Transformers	--bf16 --use_eager_attention
29/48	UCLA-AGI_Gemma-2-9B-It-SPPO-Iter3	9B	18.48	Transformers	--bf16 --use_eager_attention
29/48	bartowski_Qwen1.5-32B-Chat-exl2_5_0	32B	21.52	ExLlamav2_HF
29/48	turboderp_Llama-3.1-70B-Instruct-exl2_2.5bpw	70B	24.32	ExLlamav2_HF
29/48	microsoft_Phi-3-medium-4k-instruct	14B	27.92	Transformers
29/48	Mistral-Large-Instruct-2407-IQ2_XXS	123B	32.4	llamacpp_HF	[link]
29/48	LoneStriker_Yi-34B-Chat-8.0bpw-h8-exl2	34B	34.87	ExLlamav2_HF
29/48	34b-beta.Q8_0	34B	36.5	llamacpp_HF
29/48	Dracones_Llama-3-Lumimaid-70B-v0.1_exl2_4.5bpw	70B	41.43	ExLlamav2_HF
29/48	Llama3-ChatQA-1.5-70B.Q4_K_M	70B	42.5	llamacpp_HF	Alpaca template.
29/48	turboderp_Llama-3-70B-exl2_5.0bpw	70B	45.7	ExLlamav2_HF
29/48	turboderp_command-r-plus-103B-exl2_3.0bpw	104B	47.72	ExLlamav2_HF
29/48	daybreak-miqu-1-70b-v1.0-q5_k_m	70B	48.8	llamacpp_HF
29/48	Phi-3-medium-4k-instruct-f32	14B	55.80	llamacpp_HF	[link]
29/48	c4ai-command-r-plus.i1-IQ4_XS	104B	56.2	llamacpp_HF	[link]
29/48	Qwen1.5-110B-Chat-Q4_K_M	110B	67.2	llamacpp_HF	[link]
28/48	gemma-2-9b-it-IQ3_M	9B	4.49	llamacpp_HF	[link]
28/48	gemma-2-9b-it-Q3_K_M	9B	4.76	llamacpp_HF	[link]
28/48	gemma-2-9b-it-Q3_K_L	9B	5.13	llamacpp_HF	[link]
28/48	gemma-2-9b-it-Q3_K_L-Q8	9B	5.35	llamacpp_HF	[link]
28/48	gemma-2-9b-it-Q3_K_XL	9B	5.35	llamacpp_HF	[link]
28/48	gemma-2-9b-it-Q4_K_S	9B	5.48	llamacpp_HF	[link]
28/48	gemma-2-9b-it-Q4_K_M	9B	5.76	llamacpp_HF	[link]
28/48	gemma-2-9b-it-Q4_K_L	9B	5.98	llamacpp_HF	[link]
28/48	Phi-3-medium-128k-instruct-Q3_K_S	14B	6.06	llamacpp_HF	[link]
28/48	gemma-2-9b-it-Q5_K_M	9B	6.65	llamacpp_HF	[link]
28/48	gemma-2-9b-it-Q4_K_M-fp16	9B	6.84	llamacpp_HF	[link]
28/48	gemma-2-9b-it-Q5_K_L	9B	6.87	llamacpp_HF	[link]
28/48	Phi-3-medium-4k-instruct-Q3_K_M	14B	6.92	llamacpp_HF	[link]
28/48	gemma-2-27b-it-IQ2_M	27B	9.40	llamacpp_HF	[link]
28/48	gemma-2-9b-it-Q8_0	9B	9.83	llamacpp_HF	[link]
28/48	gemma-2-9b-it-Q8_0-f16	9B	10.69	llamacpp_HF	[link]
28/48	gemma-2-9b-it-Q8_0_L	9B	10.69	llamacpp_HF	[link]
28/48	CohereForAI_aya-expanse-32b	32B	32.30	Transformers	--load-in-8bit
28/48	gemma-2-9b-it-f32	9B	36.97	llamacpp_HF	[link]
28/48	dolphin-2.7-mixtral-8x7b.Q8_0	8x7B	49.6	llamacpp_HF
28/48	turboderp_command-r-plus-103B-exl2_3.5bpw	104B	54.2	ExLlamav2_HF
28/48	LoneStriker_Smaug-72B-v0.1-6.0bpw-h6-exl2	72B	55.82	ExLlamav2_HF
28/48	command-r-plus-Q4_K_M	104B	62.81	llamacpp_HF
28/48	turboderp_command-r-plus-103B-exl2_4.5bpw	104B	67.18	ExLlamav2_HF
27/48	gemma-2-9b-it-Q3_K_S	9B	4.34	llamacpp_HF	[link]
27/48	gemma-2-9b-it-IQ4_XS	9B	5.18	llamacpp_HF	[link]
27/48	Phi-3-medium-4k-instruct-IQ3_XS	14B	5.81	llamacpp_HF	[link]
27/48	Phi-3-medium-4k-instruct-IQ3_S	14B	6.06	llamacpp_HF	[link]
27/48	Phi-3-medium-4k-instruct-Q3_K_L	14B	7.49	llamacpp_HF	[link]
27/48	gemma-2-9b-it-Q6_K	9B	7.59	llamacpp_HF	[link]
27/48	gemma-2-9b-it-Q6_K_L	9B	7.81	llamacpp_HF	[link]
27/48	gemma-2-9b-it-Q6_K-Q8	9B	7.81	llamacpp_HF	[link]
27/48	Phi-3-medium-4k-instruct-IQ4_NL	14B	7.90	llamacpp_HF	[link]
27/48	google_gemma-2-9b-it	9B	9.24	Transformers	--load-in-8bit --bf16 --use_eager_attention
27/48	gemma-2-9b-it-Q6_K-f32	9B	10.51	llamacpp_HF	[link]
27/48	Qwen_Qwen2-7B-Instruct	7B	15.23	Transformers
27/48	01-ai_Yi-1.5-9B-Chat	9B	17.66	Transformers
27/48	Meta-Llama-3.1-70B-Instruct-IQ2_S	70B	22.2	llamacpp_HF	[link]
27/48	miqu-1-70b.q2_K	70B	25.5	llamacpp_HF
27/48	ISTA-DASLab_c4ai-command-r-plus-AQLM-2Bit-1x16	104B	31.94	Transformers
27/48	Platypus2-70B.i1-Q4_K_M	70B	41.4	llamacpp_HF	[link]
27/48	Midnight-Miqu-70B-v1.0.Q4_K_M	70B	41.7	llamacpp_HF
27/48	Llama3-ChatQA-1.5-70B.Q4_K_M	70B	42.5	llamacpp_HF	NVIDIA-ChatQA template.
27/48	mixtral-8x7b-instruct-v0.1.Q8_0	8x7B	49.6	llamacpp_HF
27/48	TheBloke_Helion-4x34B-GPTQ	4x34B	58.31	ExLlamav2_HF
26/48	Phi-3-mini-4k-instruct-Q5_K_S	3.8B	2.64	llamacpp_HF	[link]
26/48	gemma-2-9b-it-IQ3_XS	9B	4.14	llamacpp_HF	[link]
26/48	google_gemma-2-9b-it	9B	4.62	Transformers	--load-in-4bit --bf16 --use_eager_attention
26/48	Phi-3-medium-4k-instruct-IQ3_M	14B	6.47	llamacpp_HF	[link]
26/48	Phi-3-medium-4k-instruct-IQ4_XS	14B	7.47	llamacpp_HF	[link]
26/48	gemma-2-27b-it-IQ2_XS	27B	8.40	llamacpp_HF	[link]
26/48	gemma-2-27b-it-Q2_K	27B	10.45	llamacpp_HF	[link]
26/48	gemma-2-27b-it-Q2_K_L	27B	10.74	llamacpp_HF	[link]
26/48	gemma-2-27b-it-IQ3_XXS	27B	10.75	llamacpp_HF	[link]
26/48	turboderp_gemma-2-27b-it-exl2_3.0bpw	27B	13.06	ExLlamav2_HF	--no_flash_attn --no_xformers --no_sdpa
26/48	microsoft_Phi-3-small-8k-instruct	7B	14.78	Transformers
26/48	internlm_internlm2_5-7b-chat	7B	15.48	Transformers
26/48	Meta-Llama-3-70B-Instruct-IQ2_XXS	70B	19.1	llamacpp_HF
26/48	Qwen_Qwen1.5-14B-Chat	14B	28.33	Transformers
26/48	Mistral-Large-Instruct-2407-IQ1_M	123B	28.4	llamacpp_HF	[link]
26/48	LoneStriker_Yi-34B-200K-8.0bpw-h8-exl2	34B	34.84	ExLlamav2_HF
26/48	LoneStriker_dolphin-2.2-yi-34b-200k-8.0bpw-h8-exl2	34B	34.89	ExLlamav2_HF
26/48	CausalLM-RP-34B.q8_0	34B	37.0	llamacpp_HF
26/48	lzlv_70b_fp16_hf.Q5_K_M	70B	48.8	llamacpp_HF
26/48	nous-hermes-2-mixtral-8x7b-dpo.Q8_0	8x7B	49.6	llamacpp_HF
26/48	Mixtral-8x22B-Instruct-v0.1.Q4_K_M	8x22B	85.5	llamacpp_HF
25/48	Phi-3-mini-4k-instruct-Q5_K_M	3.8B	2.82	llamacpp_HF	[link]
25/48	Phi-3-mini-4k-instruct-old-Q5_K_M	3.8B	2.82	llamacpp_HF	[link]
25/48	gemma-2-9b-it-IQ2_S	9B	3.21	llamacpp_HF	[link]
25/48	gemma-2-9b-it-IQ2_M	9B	3.43	llamacpp_HF	[link]
25/48	gemma-2-9b-it-IQ3_XXS	9B	3.80	llamacpp_HF	[link]
25/48	gemma-2-9b-it-Q2_K	9B	3.81	llamacpp_HF	[link]
25/48	gemma-2-9b-it-Q2_K_L	9B	4.03	llamacpp_HF	[link]
25/48	Phi-3-medium-4k-instruct-IQ2_S	14B	4.34	llamacpp_HF	[link]
25/48	gemma-2-27b-it-IQ2_S	27B	8.65	llamacpp_HF	[link]
25/48	NousResearch_Nous-Hermes-2-SOLAR-10.7B	10.7B	21.46	Transformers
25/48	turboderp_Llama-3-70B-Instruct-exl2_2.4bpw	70B	23.47	ExLlamav2_HF
25/48	internlm_internlm2_5-20b-chat	20B	39.72	Transformers
25/48	LoneStriker_Llama-3-70B-Instruct-Gradient-524k-6.0bpw-h6-exl2	70B	54.11	ExLlamav2_HF
25/48	goliath-120b.Q4_K_M	120B	70.6	llamacpp_HF
25/48	Meta-Llama-3.1-405B-Instruct.i1-IQ1_M	405B	95.1	llamacpp_HF	[link]
24/48	Phi-3-mini-4k-instruct-IQ4_XS	3.8B	2.06	llamacpp_HF	[link]
24/48	Phi-3-mini-4k-instruct-IQ4_NL	3.8B	2.18	llamacpp_HF	[link]
24/48	Phi-3-mini-4k-instruct-old-IQ4_NL	3.8B	2.18	llamacpp_HF	[link]
24/48	Phi-3-mini-4k-instruct-old-Q5_K_S	3.8B	2.64	llamacpp_HF	[link]
24/48	Phi-3-mini-4k-instruct-Q6_K	3.8B	3.14	llamacpp_HF	[link]
24/48	Phi-3-mini-4k-instruct-old-Q6_K	3.8B	3.14	llamacpp_HF	[link]
24/48	Phi-3-mini-4k-instruct-Q8_0	3.8B	4.06	llamacpp_HF	[link]
24/48	Phi-3-mini-4k-instruct-old-Q8_0	3.8B	4.06	llamacpp_HF	[link]
24/48	Phi-3-mini-4k-instruct-old-fp16	3.8B	7.64	llamacpp_HF	[link]
24/48	Phi-3-mini-4k-instruct-fp32	3.8B	15.29	llamacpp_HF	[link]
24/48	meta-llama_Meta-Llama-3.1-8B-Instruct	8B	16.06	Transformers
24/48	Meta-Llama-3.1-70B-Instruct-IQ2_XS	70B	21.1	llamacpp_HF	[link]
24/48	upstage_SOLAR-10.7B-Instruct-v1.0	10.7B	21.46	Transformers
24/48	mistralai_Mistral-Nemo-Instruct-2407	12B	24.5	Transformers
24/48	dnhkng_RYS-Phi-3-medium-4k-instruct	14B	35.42	Transformers
24/48	MultiVerse_70B.Q4_K_M	70B	45.2	llamacpp_HF
24/48	maid-yuzu-v8-alter.Q8_0	8x7B	49.8	llamacpp_HF
24/48	LoneStriker_Llama-3-70B-Instruct-Gradient-262k-6.0bpw-h6-exl2	70B	54.27	ExLlamav2_HF
23/48	Phi-3-mini-4k-instruct-old-IQ4_XS	3.8B	2.06	llamacpp_HF	[link]
23/48	Phi-3-mini-4k-instruct-Q4_K_M	3.8B	2.39	llamacpp_HF	[link]
23/48	Phi-3-medium-128k-instruct-Q2_K	14B	5.14	llamacpp_HF	[link]
23/48	Phi-3-medium-4k-instruct-Q3_K_S	14B	6.06	llamacpp_HF	[link]
23/48	microsoft_Phi-3-mini-4k-instruct	3.8B	7.64	Transformers
23/48	Mistral-Nemo-Instruct-2407-Q8_0	12B	13.0	llamacpp_HF	[link]
23/48	Zyphra_Zamba2-7B-Instruct	7B	15.05	Transformers
23/48	bhenrym14_airoboros-3_1-yi-34b-200k	34B	34.39	Transformers	--load-in-8bit
23/48	xwin-lm-70b-v0.1.Q4_K_M	70B	41.4	llamacpp_HF
22/48	Phi-3-mini-4k-instruct-old-IQ3_XXS	3.8B	1.51	llamacpp_HF	[link]
22/48	Phi-3-mini-4k-instruct-Q4_K_S	3.8B	2.19	llamacpp_HF	[link]
22/48	Phi-3-mini-4k-instruct-old-Q4_K_M	3.8B	2.39	llamacpp_HF	[link]
22/48	Phi-3-medium-128k-instruct-IQ2_S	14B	4.34	llamacpp_HF	[link]
22/48	Meta-Llama-3-8B-Instruct-Q4_K_S	8B	4.69	llamacpp_HF
22/48	Phi-3-medium-128k-instruct-IQ2_M	14B	4.72	llamacpp_HF	[link]
22/48	Phi-3-medium-4k-instruct-IQ2_M	14B	4.72	llamacpp_HF	[link]
22/48	Phi-3-medium-4k-instruct-Q2_K	14B	5.14	llamacpp_HF	[link]
22/48	Qwen_Qwen1.5-7B-Chat	7B	15.44	Transformers
22/48	THUDM_glm-4-9b-chat	9B	18.80	Transformers
22/48	liuhaotian_llava-v1.5-13b	13B	26.09	Transformers
22/48	turboderp_command-r-v01-35B-exl2_6.0bpw	35B	32.09	ExLlamav2_HF
22/48	turboderp_command-r-plus-103B-exl2_2.5bpw	104B	41.23	ExLlamav2_HF
22/48	tulu-2-dpo-70b.Q4_K_M	70B	41.4	llamacpp_HF
22/48	wizardlm-70b-v1.0.Q4_K_M	70B	41.4	llamacpp_HF
22/48	MoMo-72B-lora-1.8.6-DPO-Q4_K_M	72B	43.8	llamacpp_HF
22/48	meraGPT_mera-mix-4x7B	4x7B	48.31	Transformers
21/48	Phi-3-mini-4k-instruct-IQ3_XXS	3.8B	1.51	llamacpp_HF	[link]
21/48	Phi-3-mini-4k-instruct-old-Q3_K_L	3.8B	2.09	llamacpp_HF	[link]
21/48	Phi-3-mini-4k-instruct-old-Q4_K_S	3.8B	2.19	llamacpp_HF	[link]
21/48	Phi-3.1-mini-4k-instruct-Q4_K_M	3.8B	2.39	llamacpp_HF	[link]
21/48	Phi-3.1-mini-4k-instruct-Q5_K_S	3.8B	2.64	llamacpp_HF	[link]
21/48	gemma-2-9b-it-IQ2_XS	9B	3.07	llamacpp_HF	[link]
21/48	Phi-3-medium-4k-instruct-IQ2_XS	14B	4.13	llamacpp_HF	[link]
21/48	Meta-Llama-3.1-8B-Instruct-Q4_K_S	8B	4.69	llamacpp_HF	[link]
21/48	Meta-Llama-3.1-8B-Instruct-Q5_K_M	8B	5.73	llamacpp_HF	[link]
21/48	Meta-Llama-3.1-8B-Instruct-Q6_K	8B	6.6	llamacpp_HF	[link]
21/48	Meta-Llama-3.1-8B-Instruct-Q6_K_L	8B	6.85	llamacpp_HF	[link]
21/48	microsoft_Phi-3-vision-128k-instruct	4.2B	8.29	Transformers
21/48	Meta-Llama-3-8B-Instruct-Q8_0	8B	8.54	llamacpp_HF
21/48	01-ai_Yi-1.5-6B-Chat	6B	12.12	Transformers
21/48	microsoft_Phi-3-small-128k-instruct	7B	14.78	Transformers
21/48	gustavecortal_oneirogen-7B	7B	15.23	Transformers
21/48	NurtureAI_Meta-Llama-3-8B-Instruct-64k	8B	16.06	Transformers
21/48	Undi95_Meta-Llama-3-8B-Instruct-hf	8B	16.06	Transformers
21/48	Salesforce_SFR-Iterative-DPO-LLaMA-3-8B-R	8B	16.06	Transformers
21/48	Meta-Llama-3-8B-Instruct-fp16	8B	16.1	llamacpp_HF
21/48	c4ai-command-r-v01-Q8_0	35B	37.2	llamacpp_HF
21/48	internlm_internlm2-chat-20b	20B	39.72	Transformers
21/48	falcon-180b.Q4_K_M	180B	108.6	llamacpp_HF
20/48	gemma-2-2b-it-Q4_K_S	2.6B	1.64	llamacpp_HF	[link]
20/48	Phi-3-mini-4k-instruct-Q3_K_M	3.8B	1.96	llamacpp_HF	[link]
20/48	Phi-3-mini-4k-instruct-Q3_K_L	3.8B	2.09	llamacpp_HF	[link]
20/48	Phi-3.1-mini-4k-instruct-Q4_K_S	3.8B	2.19	llamacpp_HF	[link]
20/48	Phi-3.1-mini-4k-instruct-Q5_K_M	3.8B	2.82	llamacpp_HF	[link]
20/48	Phi-3.1-mini-4k-instruct-Q6_K	3.8B	3.14	llamacpp_HF	[link]
20/48	Phi-3.1-mini-4k-instruct-Q6_K_L	3.8B	3.18	llamacpp_HF	[link]
20/48	Phi-3-medium-128k-instruct-IQ2_XS	14B	4.13	llamacpp_HF	[link]
20/48	Meta-Llama-3.1-8B-Instruct-Q4_K_M	8B	4.92	llamacpp_HF	[link]
20/48	Meta-Llama-3.1-8B-Instruct-Q4_K_L	8B	5.31	llamacpp_HF	[link]
20/48	Meta-Llama-3.1-8B-Instruct-Q5_K_L	8B	6.06	llamacpp_HF	[link]
20/48	Meta-Llama-3-8B-Instruct-Q6_K	8B	6.6	llamacpp_HF
20/48	TheBloke_llava-v1.5-13B-GPTQ	13B	7.26	ExLlamav2_HF
20/48	microsoft_Phi-3-mini-4k-instruct-20240701	3.8B	7.64	Transformers
20/48	openchat_openchat_3.5	7B	14.48	Transformers
20/48	Weyaxi_Einstein-v6.1-Llama3-8B	8B	16.06	Transformers
20/48	BAAI_Bunny-Llama-3-8B-V	8B	16.96	Transformers
20/48	CohereForAI_aya-23-35B	35B	34.98	Transformers	--load-in-8bit
20/48	llama-2-70b-chat.Q4_K_M	70B	41.4	llamacpp_HF
20/48	Ein-72B-v0.1-full.Q4_K_M	72B	45.2	llamacpp_HF
19/48	Phi-3.1-mini-128k-instruct-IQ3_XXS	3.8B	1.51	llamacpp_HF	[link]
19/48	Phi-3.1-mini-128k-instruct-IQ3_XS	3.8B	1.63	llamacpp_HF	[link]
19/48	gemma-2-2b-it-Q4_K_M	2.6B	1.71	llamacpp_HF	[link]
19/48	Phi-3-mini-4k-instruct-IQ3_M	3.8B	1.86	llamacpp_HF	[link]
19/48	Phi-3-mini-4k-instruct-old-IQ3_M	3.8B	1.86	llamacpp_HF	[link]
19/48	gemma-2-2b-it-Q5_K_S	2.6B	1.88	llamacpp_HF	[link]
19/48	Phi-3.1-mini-4k-instruct-Q3_K_M	3.8B	1.96	llamacpp_HF	[link]
19/48	Phi-3-mini-4k-instruct-old-Q3_K_M	3.8B	1.96	llamacpp_HF	[link]
19/48	Phi-3.1-mini-4k-instruct-IQ4_XS	3.8B	2.06	llamacpp_HF	[link]
19/48	Phi-3.1-mini-128k-instruct-Q3_K_L	3.8B	2.09	llamacpp_HF	[link]
19/48	Phi-3.1-mini-4k-instruct-Q3_K_L	3.8B	2.09	llamacpp_HF	[link]
19/48	Phi-3.1-mini-128k-instruct-Q3_K_XL	3.8B	2.17	llamacpp_HF	[link]
19/48	Phi-3.1-mini-4k-instruct-Q3_K_XL	3.8B	2.17	llamacpp_HF	[link]
19/48	Phi-3.1-mini-128k-instruct-Q4_K_L	3.8B	2.47	llamacpp_HF	[link]
19/48	Phi-3.1-mini-4k-instruct-Q4_K_L	3.8B	2.47	llamacpp_HF	[link]
19/48	Phi-3.1-mini-128k-instruct-Q5_K_S	3.8B	2.64	llamacpp_HF	[link]
19/48	Phi-3.1-mini-128k-instruct-Q5_K_M	3.8B	2.82	llamacpp_HF	[link]
19/48	Phi-3.1-mini-128k-instruct-Q5_K_L	3.8B	2.88	llamacpp_HF	[link]
19/48	Phi-3.1-mini-4k-instruct-Q5_K_L	3.8B	2.88	llamacpp_HF	[link]
19/48	Meta-Llama-3-8B-Instruct-IQ3_S	8B	3.68	llamacpp_HF	[link]
19/48	Phi-3.1-mini-4k-instruct-Q8_0	3.8B	4.06	llamacpp_HF	[link]
19/48	mistral-7b-instruct-v0.2.Q4_K_S	7B	4.14	llamacpp_HF
19/48	Meta-Llama-3-8B-Instruct-IQ4_XS	8B	4.45	llamacpp_HF
19/48	Meta-Llama-3-8B-Instruct-IQ4_NL	8B	4.68	llamacpp_HF
19/48	Meta-Llama-3-8B-Instruct-Q4_K_M	8B	4.92	llamacpp_HF
19/48	Meta-Llama-3-8B-Instruct-Q5_K_S	8B	5.6	llamacpp_HF
19/48	Meta-Llama-3-8B-Instruct-Q5_K_M	8B	5.73	llamacpp_HF
19/48	microsoft_Phi-3-mini-128k-instruct	3.8B	7.64	Transformers
19/48	microsoft_Phi-3.5-vision-instruct	3.8B	8.29	Transformers
19/48	Nexusflow_Starling-LM-7B-beta	7B	14.48	Transformers
19/48	NousResearch_Hermes-2-Pro-Mistral-7B	7B	14.48	Transformers
19/48	Phi-3.1-mini-4k-instruct-f32	3.8B	15.29	llamacpp_HF	[link]
19/48	internlm_internlm2-chat-7b	7B	15.48	Transformers
19/48	lightblue_suzume-llama-3-8B-multilingual	8B	16.06	Transformers
19/48	UCLA-AGI_Llama-3-Instruct-8B-SPPO-Iter3	8B	16.06	Transformers
19/48	ai21labs_Jamba-v0.1	52B	25.79	Transformers	--load-in-4bit
19/48	Meta-Llama-3.1-8B-Instruct-f32	8B	32.1	llamacpp_HF	[link]
19/48	internlm_internlm2-chat-20b-sft	20B	39.72	Transformers
19/48	llama-2-70b.Q5_K_M	70B	48.8	llamacpp_HF
19/48	DeepSeek-Coder-V2-Instruct-IQ2_XS	236B	68.7	llamacpp_HF	[link]
19/48	zephyr-orpo-141b-A35b-v0.1.Q4_K_M	141B	85.59	llamacpp_HF
18/48	gemma-2-2b-it-IQ4_XS	2.6B	1.57	llamacpp_HF	[link]
18/48	Phi-3.1-mini-4k-instruct-IQ3_XS	3.8B	1.63	llamacpp_HF	[link]
18/48	Phi-3-mini-4k-instruct-IQ3_S	3.8B	1.68	llamacpp_HF	[link]
18/48	Phi-3-mini-4k-instruct-Q3_K_S	3.8B	1.68	llamacpp_HF	[link]
18/48	Phi-3-mini-4k-instruct-old-IQ3_S	3.8B	1.68	llamacpp_HF	[link]
18/48	Phi-3-mini-4k-instruct-old-Q3_K_S	3.8B	1.68	llamacpp_HF	[link]
18/48	Phi-3.1-mini-128k-instruct-IQ3_M	3.8B	1.86	llamacpp_HF	[link]
18/48	Phi-3.1-mini-4k-instruct-IQ3_M	3.8B	1.86	llamacpp_HF	[link]
18/48	gemma-2-2b-it-Q5_K_M	2.6B	1.92	llamacpp_HF	[link]
18/48	Phi-3.1-mini-128k-instruct-Q3_K_M	3.8B	1.96	llamacpp_HF	[link]
18/48	gemma-2-2b-it-Q6_K	2.6B	2.15	llamacpp_HF	[link]
18/48	Phi-3.1-mini-128k-instruct-Q4_K_S	3.8B	2.19	llamacpp_HF	[link]
18/48	gemma-2-2b-it-Q6_K_L	2.6B	2.29	llamacpp_HF	[link]
18/48	Phi-3.1-mini-128k-instruct-Q4_K_M	3.8B	2.39	llamacpp_HF	[link]
18/48	gemma-2-2b-it-Q8_0	2.6B	2.78	llamacpp_HF	[link]
18/48	Phi-3.1-mini-128k-instruct-Q6_K	3.8B	3.14	llamacpp_HF	[link]
18/48	Phi-3.1-mini-128k-instruct-Q6_K_L	3.8B	3.18	llamacpp_HF	[link]
18/48	Meta-Llama-3-8B-Instruct-Q3_K_M	8B	4.02	llamacpp_HF
18/48	Meta-Llama-3-8B-Instruct-Q3_K_L	8B	4.32	llamacpp_HF
18/48	google_gemma-2-2b-it	2B	5.23	Transformers
18/48	microsoft_Phi-3-mini-128k-instruct	3.8B	7.64	Transformers	--load-in-8bit
18/48	microsoft_Phi-3.5-mini-instruct	3.8B	8.29	Transformers
18/48	Meta-Llama-3.1-8B-Instruct-Q8_0	8B	8.54	llamacpp_HF	[link]
18/48	gemma-2-2b-it-f32	2.6B	10.46	llamacpp_HF	[link]
18/48	mistralai_Mistral-7B-Instruct-v0.2	7B	14.48	Transformers
18/48	jieliu_Storm-7B	7B	14.48	Transformers
18/48	failspy_kappa-3-phi-abliterated	3.8B	15.28	Transformers
18/48	Phi-3.1-mini-128k-instruct-f32	3.8B	15.29	llamacpp_HF	[link]
18/48	turboderp_llama3-turbcat-instruct-8b	8B	16.06	Transformers
18/48	Meta-Llama-3-70B-Instruct-IQ1_M	70B	16.8	llamacpp_HF
18/48	Orenguteng_Lexi-Llama-3-8B-Uncensored	8B	17.67	Transformers
18/48	Meta-Llama-3.1-70B-Instruct-IQ2_XXS	70B	19.1	llamacpp_HF	[link]
18/48	LoneStriker_Nous-Capybara-34B-4.65bpw-h6-exl2	34B	20.76	ExLlamav2_HF
18/48	Qwen_Qwen1.5-MoE-A2.7B-Chat	14.3B	28.63	Transformers
18/48	TheProfessor-155b.i1-IQ3_XS	155B	63.2	llamacpp_HF	[link]
17/48	gemma-2-2b-it-IQ3_M	2.6B	1.39	llamacpp_HF	[link]
17/48	Phi-3-mini-4k-instruct-v0.3-IQ3_XXS	3.8B	1.51	llamacpp_HF	[link]
17/48	gemma-2-2b-it-Q3_K_L	2.6B	1.55	llamacpp_HF	[link]
17/48	Phi-3-mini-4k-instruct-v0.3-Q3_K_S	3.8B	1.68	llamacpp_HF	[link]
17/48	microsoft_Phi-3-mini-128k-instruct	3.8B	1.91	Transformers	--load-in-4bit
17/48	turboderp_Phi-3-mini-128k-instruct-exl2_5.0bpw	3.8B	2.54	ExLlamav2_HF
17/48	Phi-3-mini-4k-instruct-v0.3-Q5_K_S	3.8B	2.64	llamacpp_HF	[link]
17/48	turboderp_Phi-3-mini-128k-instruct-exl2_6.0bpw	3.8B	2.99	ExLlamav2_HF
17/48	Phi-3.1-mini-128k-instruct-Q8_0	3.8B	4.06	llamacpp_HF	[link]
17/48	Meta-Llama-3.1-8B-Instruct-Q3_K_L	8B	4.32	llamacpp_HF	[link]
17/48	Meta-Llama-3.1-8B-Instruct-IQ4_XS	8B	4.45	llamacpp_HF	[link]
17/48	Meta-Llama-3.1-8B-Instruct-Q5_K_S	8B	5.6	llamacpp_HF	[link]
17/48	amazingvince_Not-WizardLM-2-7B	7B	14.48	Transformers
17/48	Undi95_Toppy-M-7B	7B	14.48	Transformers
17/48	internlm_internlm2-chat-7b-sft	7B	15.48	Transformers
17/48	mzbac_llama-3-8B-Instruct-function-calling	8B	16.06	Transformers
17/48	openchat_openchat-3.6-8b-20240522	8B	16.06	Transformers
17/48	kubernetes-bad_Mistral-7B-Instruct-v0.3	7B	28.99	Transformers
17/48	ZeusLabs_L3-Aethora-15B-V2	15B	30.02	Transformers
17/48	ggml-alpaca-dragon-72b-v1-q4_k_m	72B	43.8	llamacpp_HF
17/48	DeepSeek-V2-Chat-0628-IQ2_XS	236B	68.7	llamacpp_HF	[link]
17/48	Meta-Llama-3.1-405B-Instruct.i1-IQ1_S	405B	86.8	llamacpp_HF	[link]
17/48	grok-1-IQ2_XS	314B	93.31	llamacpp_HF	[link]
16/48	Phi-3.1-mini-4k-instruct-Q3_K_S	3.8B	1.68	llamacpp_HF	[link]
16/48	Phi-3.1-mini-128k-instruct-IQ4_XS	3.8B	2.06	llamacpp_HF	[link]
16/48	Phi-3-mini-4k-instruct-v0.3-Q4_K_S	3.8B	2.19	llamacpp_HF	[link]
16/48	Meta-Llama-3-8B-Instruct-IQ3_M	8B	3.78	llamacpp_HF
16/48	Meta-Llama-3.1-8B-Instruct-Q3_K_M	8B	4.02	llamacpp_HF	[link]
16/48	TheBloke_Mistral-7B-Instruct-v0.2-GPTQ	7B	4.16	ExLlamav2_HF
16/48	Meta-Llama-3.1-8B-Instruct-Q3_K_XL	8B	4.78	llamacpp_HF	[link]
16/48	microsoft_Phi-3-mini-128k-instruct-20240701	3.8B	7.64	Transformers
16/48	mixtral-8x7b-instruct-v0.1.Q2_K	8x7B	15.6	llamacpp_HF
15/48	Phi-3-mini-4k-instruct-IQ3_XS	3.8B	1.63	llamacpp_HF	[link]
15/48	Phi-3-mini-4k-instruct-old-IQ3_XS	3.8B	1.63	llamacpp_HF	[link]
15/48	Phi-3.1-mini-128k-instruct-Q3_K_S	3.8B	1.68	llamacpp_HF	[link]
15/48	Phi-3-mini-4k-instruct-v0.3-Q3_K_L	3.8B	2.09	llamacpp_HF	[link]
15/48	Phi-3-mini-4k-instruct-v0.3-Q4_K_M	3.8B	2.39	llamacpp_HF	[link]
15/48	Phi-3-mini-4k-instruct-v0.3-Q5_K_M	3.8B	2.82	llamacpp_HF	[link]
15/48	Meta-Llama-3-8B-Instruct-IQ3_XS	8B	3.52	llamacpp_HF
15/48	Meta-Llama-3.1-8B-Instruct-Q3_K_S	8B	3.66	llamacpp_HF	[link]
15/48	hjhj3168_Llama-3-8b-Orthogonalized-exl2	8B	6.7	ExLlamav2_HF
15/48	cognitivecomputations_dolphin-2.9-llama3-8b	8B	16.06	Transformers
15/48	xtuner_llava-llama-3-8b-v1_1	8B	16.06	Transformers
15/48	NousResearch_Hermes-2-Pro-Llama-3-8B	8B	16.06	Transformers
15/48	CohereForAI_c4ai-command-r-v01-4bit	35B	22.69	Transformers
14/48	turboderp_Phi-3-mini-128k-instruct-exl2_4.0bpw	3.8B	2.09	ExLlamav2_HF
14/48	Meta-Llama-3-8B-Instruct-IQ3_XXS	8B	3.27	llamacpp_HF
14/48	Meta-Llama-3.1-8B-Instruct-IQ3_XS	8B	3.52	llamacpp_HF	[link]
14/48	Meta-Llama-3-8B-Instruct-Q3_K_S	8B	3.66	llamacpp_HF
14/48	Meta-Llama-3.1-8B-Instruct-IQ3_M	8B	3.78	llamacpp_HF	[link]
14/48	mattshumer_Llama-3-8B-16K	8B	16.06	Transformers
14/48	nvidia_ChatQA-1.5-8B	8B	16.06	Transformers	Alpaca template.
14/48	Gryphe_MythoMax-L2-13b	13B	26.03	Transformers
14/48	microsoft_Orca-2-13b	13B	52.06	Transformers
14/48	Undi95_ReMM-SLERP-L2-13B	13B	52.06	Transformers
13/48	Phi-3-mini-4k-instruct-v0.3-Q3_K_M	3.8B	1.96	llamacpp_HF	[link]
13/48	Phi-3-mini-4k-instruct-v0.3-Q6_K	3.8B	3.14	llamacpp_HF	[link]
13/48	Phi-3-medium-4k-instruct-IQ2_XXS	14B	3.72	llamacpp_HF	[link]
13/48	gradientai_Llama-3-8B-Instruct-262k	8B	16.06	Transformers
13/48	gradientai_Llama-3-8B-Instruct-Gradient-1048k	8B	16.06	Transformers
13/48	nvidia_ChatQA-1.5-8B	8B	16.06	Transformers	NVIDIA-ChatQA template.
13/48	alpindale_gemma-7b-it	7B	17.08	Transformers
13/48	internlm_internlm2-wqx-20b	20B	39.72	Transformers
13/48	GeorgiaTechResearchInstitute_galactica-30b-evol-instruct-70k	30B	59.95	Transformers	GALACTICA template.
13/48	turboderp_dbrx-instruct-exl2_3.75bpw	132B	62.83	ExLlamav2_HF	Without the "You are DBRX..." system prompt.
12/48	Phi-3.1-mini-128k-instruct-Q2_K	3.8B	1.42	llamacpp_HF	[link]
12/48	Phi-3.1-mini-128k-instruct-Q2_K_L	3.8B	1.51	llamacpp_HF	[link]
12/48	Phi-3-mini-4k-instruct-v0.3-Q8_0	3.8B	4.06	llamacpp_HF	[link]
12/48	ISTA-DASLab_Meta-Llama-3-8B-Instruct-AQLM-2Bit-1x16	8B	4.08	Transformers
12/48	HuggingFaceH4_zephyr-7b-beta	7B	14.48	Transformers
12/48	Phi-3-mini-4k-instruct-v0.3-f32	3.8B	15.29	llamacpp_HF	[link]
12/48	Meta-Llama-3-70B-Instruct-IQ1_S	70B	15.3	llamacpp_HF
12/48	NousResearch_Llama-2-13b-chat-hf	13B	26.03	Transformers
12/48	mistralai_mathstral-7B-v0.1	7B	28.99	Transformers
12/48	llama-65b.Q5_K_M	65B	46.2	llamacpp_HF
11/48	Phi-3-mini-4k-instruct-Q2_K	3.8B	1.42	llamacpp_HF	[link]
11/48	Phi-3-mini-4k-instruct-v0.3-IQ3_M	3.8B	1.86	llamacpp_HF	[link]
11/48	Phi-3-mini-4k-instruct-v0.3-IQ4_XS	3.8B	2.06	llamacpp_HF	[link]
11/48	Meta-Llama-3-8B-Instruct-IQ2_M	8B	2.95	llamacpp_HF
11/48	Meta-Llama-3.1-8B-Instruct-IQ2_M	8B	2.95	llamacpp_HF	[link]
11/48	Meta-Llama-3-8B-Instruct-Q2_K	8B	3.18	llamacpp_HF
11/48	Phi-3-medium-128k-instruct-IQ2_XXS	14B	3.72	llamacpp_HF	[link]
11/48	meta-llama_Llama-3.2-3B-Instruct	3B	6.43	Transformers
11/48	mlabonne_phixtral-2x2_8	2x2.8B	8.92	Transformers
11/48	Meta-Llama-3.1-70B-Instruct-IQ1_M	70B	16.8	llamacpp_HF	[link]
11/48	mlfoundations_tabula-8b	8B	32.12	Transformers
11/48	nyunai_nyun-c2-llama3-56B	56B	55.62	Transformers	--load-in-8bit
10/48	ISTA-DASLab_c4ai-command-r-v01-AQLM-2Bit-1x16	35B	12.72	Transformers
10/48	facebook_galactica-30b	30B	60.66	Transformers
9/48	Phi-3-mini-4k-instruct-old-Q2_K	3.8B	1.42	llamacpp_HF	[link]
9/48	Qwen_Qwen2-1.5B-Instruct	1.5B	3.09	Transformers
9/48	Meta-Llama-3.1-8B-Instruct-Q2_K_L	8B	3.69	llamacpp_HF	[link]
9/48	microsoft_phi-2	2.7B	5.56	Transformers
9/48	TheBloke_vicuna-33B-GPTQ	33B	16.94	ExLlamav2_HF
8/48	Phi-3-mini-4k-instruct-v0.3-Q2_K	3.8B	1.42	llamacpp_HF	[link]
8/48	Meta-Llama-3-8B-Instruct-IQ2_S	8B	2.76	llamacpp_HF
8/48	ISTA-DASLab_Meta-Llama-3-8B-Instruct-AQLM-2Bit-1x16	8B	4.08	Transformers
8/48	NousResearch_Llama-2-7b-chat-hf	7B	13.48	Transformers
8/48	mistralai_Mistral-7B-Instruct-v0.1	7B	14.48	Transformers
8/48	NousResearch_Nous-Capybara-7B-V1.9	7B	14.48	Transformers
8/48	gradientai_Llama-3-8B-Instruct-Gradient-1048k	8B	16.06	Transformers	Revision of 2024/05/04.
7/48	Phi-3.1-mini-4k-instruct-Q2_K	3.8B	1.42	llamacpp_HF	[link]
7/48	Phi-3-mini-4k-instruct-v0.3-IQ3_XS	3.8B	1.63	llamacpp_HF	[link]
7/48	Meta-Llama-3.1-8B-Instruct-Q2_K	8B	3.18	llamacpp_HF	[link]
7/48	h2oai_h2o-danube3-4b-chat	4B	7.92	Transformers
7/48	tiiuae_falcon-11B	11B	22.21	Transformers
7/48	GeorgiaTechResearchInstitute_galactica-30b-evol-instruct-70k	30B	59.95	Transformers	Alpaca template.
6/48	Phi-3-mini-4k-instruct-IQ2_S	3.8B	1.22	llamacpp_HF	[link]
6/48	Phi-3.1-mini-4k-instruct-Q2_K_L	3.8B	1.51	llamacpp_HF	[link]
6/48	meta-llama_Meta-Llama-3.1-8B	8B	16.06	Transformers
6/48	tiiuae_falcon-40b-instruct	40B	41.84	Transformers	--load-in-8bit; falcon-180B-chat instruction template.
5/48	Phi-3-mini-4k-instruct-old-IQ2_S	3.8B	1.22	llamacpp_HF	[link]
5/48	Meta-Llama-3-8B-Instruct-IQ2_XS	8B	2.61	llamacpp_HF
5/48	Phi-3-medium-128k-instruct-IQ1_S	14B	2.96	llamacpp_HF	[link]
5/48	internlm_internlm2-chat-1_8b-sft	1.8B	3.78	Transformers
5/48	TheBloke_Llama-2-13B-GPTQ	13B	7.26	ExLlamav2_HF
5/48	LoneStriker_deepseek-coder-33b-instruct-6.0bpw-h6-exl2	33B	25.32	ExLlamav2_HF
5/48	NousResearch_Llama-2-13b-hf	13B	26.03	Transformers
5/48	unsloth_llama-3-70b-bnb-4bit	70B	39.52	Transformers
4/48	Phi-3-mini-4k-instruct-IQ2_M	3.8B	1.32	llamacpp_HF	[link]
4/48	turboderp_Phi-3-mini-128k-instruct-exl2_3.0bpw	3.8B	1.63	ExLlamav2_HF
4/48	internlm_internlm2-chat-1_8b	1.8B	3.78	Transformers
4/48	TheBloke_deepseek-coder-33B-instruct-AWQ	33B	18.01	AutoAWQ
3/48	Phi-3-mini-4k-instruct-IQ2_XXS	3.8B	1.04	llamacpp_HF	[link]
3/48	Phi-3-mini-4k-instruct-old-IQ2_XXS	3.8B	1.04	llamacpp_HF	[link]
3/48	Phi-3.1-mini-4k-instruct-IQ2_M	3.8B	1.32	llamacpp_HF	[link]
3/48	Phi-3-mini-4k-instruct-old-IQ2_M	3.8B	1.32	llamacpp_HF	[link]
3/48	Phi-3-medium-128k-instruct-IQ1_M	14B	3.24	llamacpp_HF	[link]
3/48	TheBloke_Llama-2-7B-GPTQ	7B	3.9	ExLlamav2_HF
3/48	turboderp_dbrx-instruct-exl2_3.75bpw	132B	62.83	ExLlamav2_HF
2/48	Phi-3-mini-4k-instruct-v0.3-IQ2_S	3.8B	1.22	llamacpp_HF	[link]
2/48	Meta-Llama-3-8B-Instruct-IQ2_XXS	8B	2.4	llamacpp_HF
2/48	Phi-3-medium-4k-instruct-IQ1_M	14B	3.24	llamacpp_HF	[link]
2/48	facebook_galactica-6.7b	6.7B	13.72	Transformers
1/48	Phi-3.1-mini-128k-instruct-IQ2_XS	3.8B	1.15	llamacpp_HF	[link]
1/48	Phi-3-mini-4k-instruct-v0.3-IQ2_XS	3.8B	1.15	llamacpp_HF	[link]
1/48	Phi-3-mini-4k-instruct-IQ2_XS	3.8B	1.15	llamacpp_HF	[link]
1/48	Phi-3-mini-4k-instruct-old-IQ2_XS	3.8B	1.15	llamacpp_HF	[link]
1/48	Phi-3.1-mini-128k-instruct-IQ2_S	3.8B	1.22	llamacpp_HF	[link]
1/48	Phi-3.1-mini-128k-instruct-IQ2_M	3.8B	1.32	llamacpp_HF	[link]
1/48	turboderp_Phi-3-mini-128k-instruct-exl2_2.5bpw	3.8B	1.41	ExLlamav2_HF
1/48	Phi-3-medium-4k-instruct-IQ1_S	14B	2.96	llamacpp_HF	[link]
1/48	bartowski_CodeQwen1.5-7B-Chat-exl2_8_0	7B	7.63	ExLlamav2_HF
1/48	NousResearch_Llama-2-7b-hf	7B	13.48	Transformers
1/48	Qwen_CodeQwen1.5-7B-Chat	7B	14.5	Transformers
0/48	facebook_galactica-125m	0.125B	0.25	Transformers
0/48	openai-community_gpt2	0.124B	0.55	Transformers
0/48	Phi-3-mini-4k-instruct-IQ1_S	3.8B	0.84	llamacpp_HF	[link]
0/48	Phi-3-mini-4k-instruct-old-IQ1_S	3.8B	0.84	llamacpp_HF	[link]
0/48	Phi-3-mini-4k-instruct-IQ1_M	3.8B	0.92	llamacpp_HF	[link]
0/48	Phi-3-mini-4k-instruct-old-IQ1_M	3.8B	0.92	llamacpp_HF	[link]
0/48	Qwen_Qwen2-0.5B-Instruct	0.5B	0.99	Transformers
0/48	h2oai_h2o-danube3-500m-chat	0.5B	1.03	Transformers
0/48	Phi-3-mini-4k-instruct-v0.3-IQ2_M	3.8B	1.32	llamacpp_HF	[link]
0/48	openai-community_gpt2-medium	0.355B	1.52	Transformers
0/48	Meta-Llama-3-8B-Instruct-IQ1_S	8B	2.02	llamacpp_HF
0/48	Meta-Llama-3-8B-Instruct-IQ1_M	8B	2.16	llamacpp_HF
0/48	TinyLlama_TinyLlama-1.1B-Chat-v1.0	1.1B	2.2	Transformers
0/48	ISTA-DASLab_Llama-2-7b-AQLM-2Bit-1x16-hf	7B	2.38	Transformers
0/48	meta-llama_Llama-3.2-1B-Instruct	1B	2.47	Transformers
0/48	facebook_galactica-1.3b	1.3B	2.63	Transformers
0/48	openai-community_gpt2-large	0.774B	3.25	Transformers
0/48	EleutherAI_gpt-neo-1.3B	1.3B	5.31	Transformers
0/48	openai-community_gpt2-xl	1.5B	6.43	Transformers
0/48	EleutherAI_gpt-neo-2.7B	2.7B	10.67	Transformers
0/48	facebook_opt-6.7b	6.7B	13.32	Transformers
0/48	gpt4chan_model_float16	6B	24.21	Transformers
0/48	EleutherAI_gpt-j-6b	6B	24.21	Transformers
0/48	facebook_opt-13b	13B	25.71	Transformers
0/48	EleutherAI_gpt-neox-20b	20B	41.29	Transformers
0/48	facebook_opt-30b	30B	59.95	Transformers

Updates

2024/11/18

Mistral-Large-Instruct-2411-IQ4_XS (123B) - 35/48

2024/10/24

THUDM_glm-4-9b-chat (9B) - 22/48
CohereForAI_aya-expanse-32b (32B) - 28/48

2024/10/16

Llama-3.1-Nemotron-70B-Instruct-HF-Q4_K_M (70B) - 34/48

2024/10/14

Zyphra_Zamba2-7B-Instruct (7B) - 23/48

2024/09/27

Reflection-Llama-3.1-70B-Q4_K_M (70B) - 31/48
meta-llama_Llama-3.2-1B-Instruct (1B) - 0/48
Qwen_Qwen2.5-7B-Instruct (7B) - 30/48
meta-llama_Llama-3.2-3B-Instruct (3B) - 11/48
Qwen2.5-32B-Instruct-Q8_0 (32B) - 34/48
Qwen2.5-72B-Instruct-Q4_K_M (72B) - 37/48

2024/08/20

microsoft_Phi-3.5-vision-instruct (3.8B) - 19/48
microsoft_Phi-3.5-mini-instruct (3.8B) - 18/48
microsoft_Phi-3.5-MoE-instruct (16x3.8B) - 30/48
microsoft_Phi-3.5-MoE-instruct (16x3.8B) - 33/48

2024/08/10

dnhkng_RYS-Gemma-2-27b-it (27B) - 32/48
dnhkng_RYS-Phi-3-medium-4k-instruct (14B) - 24/48

2024/08/08

RYS-XLarge-Q4_K_M (72B) - 35/48

2024/08/05

internlm_internlm2_5-20b-chat (20B) - 25/48

2024/08/04

gemma-2-2b-it-f32 (2.6B) - 18/48
gemma-2-2b-it-IQ3_M (2.6B) - 17/48
gemma-2-2b-it-IQ4_XS (2.6B) - 18/48
gemma-2-2b-it-Q3_K_L (2.6B) - 17/48
gemma-2-2b-it-Q4_K_M (2.6B) - 19/48
gemma-2-2b-it-Q4_K_S (2.6B) - 20/48
gemma-2-2b-it-Q5_K_M (2.6B) - 18/48
gemma-2-2b-it-Q5_K_S (2.6B) - 19/48
gemma-2-2b-it-Q6_K (2.6B) - 18/48
gemma-2-2b-it-Q6_K_L (2.6B) - 18/48
gemma-2-2b-it-Q8_0 (2.6B) - 18/48
gemma-2-9b-it-f32 (9B) - 28/48
gemma-2-9b-it-IQ2_M (9B) - 25/48
gemma-2-9b-it-IQ2_S (9B) - 25/48
gemma-2-9b-it-IQ2_XS (9B) - 21/48
gemma-2-9b-it-IQ3_M (9B) - 28/48
gemma-2-9b-it-IQ3_XS (9B) - 26/48
gemma-2-9b-it-IQ3_XXS (9B) - 25/48
gemma-2-9b-it-IQ4_XS (9B) - 27/48
gemma-2-9b-it-Q2_K (9B) - 25/48
gemma-2-9b-it-Q2_K_L (9B) - 25/48
gemma-2-9b-it-Q3_K_L (9B) - 28/48
gemma-2-9b-it-Q3_K_L-Q8 (9B) - 28/48
gemma-2-9b-it-Q3_K_M (9B) - 28/48
gemma-2-9b-it-Q3_K_S (9B) - 27/48
gemma-2-9b-it-Q3_K_XL (9B) - 28/48
gemma-2-9b-it-Q4_K_L (9B) - 28/48
gemma-2-9b-it-Q4_K_M-fp16 (9B) - 28/48
gemma-2-9b-it-Q4_K_M (9B) - 28/48
gemma-2-9b-it-Q4_K_S (9B) - 28/48
gemma-2-9b-it-Q5_K_L (9B) - 28/48
gemma-2-9b-it-Q5_K_M (9B) - 28/48
gemma-2-9b-it-Q5_K_S (9B) - 29/48
gemma-2-9b-it-Q6_K-f32 (9B) - 27/48
gemma-2-9b-it-Q6_K (9B) - 27/48
gemma-2-9b-it-Q6_K_L (9B) - 27/48
gemma-2-9b-it-Q6_K-Q8 (9B) - 27/48
gemma-2-9b-it-Q8_0-f16 (9B) - 28/48
gemma-2-9b-it-Q8_0 (9B) - 28/48
gemma-2-9b-it-Q8_0_L (9B) - 28/48
Phi-3.1-mini-128k-instruct-f32 (3.8B) - 18/48
Phi-3.1-mini-128k-instruct-IQ2_M (3.8B) - 1/48
Phi-3.1-mini-128k-instruct-IQ3_M (3.8B) - 18/48
Phi-3.1-mini-128k-instruct-IQ3_XS (3.8B) - 19/48
Phi-3.1-mini-128k-instruct-IQ4_XS (3.8B) - 16/48
Phi-3.1-mini-128k-instruct-Q2_K (3.8B) - 12/48
Phi-3.1-mini-128k-instruct-Q2_K_L (3.8B) - 12/48
Phi-3.1-mini-128k-instruct-Q3_K_L (3.8B) - 19/48
Phi-3.1-mini-128k-instruct-Q3_K_M (3.8B) - 18/48
Phi-3.1-mini-128k-instruct-Q3_K_S (3.8B) - 15/48
Phi-3.1-mini-128k-instruct-Q3_K_XL (3.8B) - 19/48
Phi-3.1-mini-128k-instruct-Q4_K_L (3.8B) - 19/48
Phi-3.1-mini-128k-instruct-Q4_K_M (3.8B) - 18/48
Phi-3.1-mini-128k-instruct-Q4_K_S (3.8B) - 18/48
Phi-3.1-mini-128k-instruct-Q5_K_L (3.8B) - 19/48
Phi-3.1-mini-128k-instruct-Q5_K_M (3.8B) - 19/48
Phi-3.1-mini-128k-instruct-Q5_K_S (3.8B) - 19/48
Phi-3.1-mini-128k-instruct-Q6_K (3.8B) - 18/48
Phi-3.1-mini-128k-instruct-Q6_K_L (3.8B) - 18/48
Phi-3.1-mini-128k-instruct-Q8_0 (3.8B) - 17/48
Phi-3.1-mini-4k-instruct-f32 (3.8B) - 19/48
Phi-3.1-mini-4k-instruct-IQ2_M (3.8B) - 3/48
Phi-3.1-mini-4k-instruct-IQ3_M (3.8B) - 18/48
Phi-3.1-mini-4k-instruct-IQ3_XS (3.8B) - 18/48
Phi-3.1-mini-4k-instruct-IQ4_XS (3.8B) - 19/48
Phi-3.1-mini-4k-instruct-Q2_K (3.8B) - 7/48
Phi-3.1-mini-4k-instruct-Q2_K_L (3.8B) - 6/48
Phi-3.1-mini-4k-instruct-Q3_K_L (3.8B) - 19/48
Phi-3.1-mini-4k-instruct-Q3_K_M (3.8B) - 19/48
Phi-3.1-mini-4k-instruct-Q3_K_S (3.8B) - 16/48
Phi-3.1-mini-4k-instruct-Q3_K_XL (3.8B) - 19/48
Phi-3.1-mini-4k-instruct-Q4_K_L (3.8B) - 19/48
Phi-3.1-mini-4k-instruct-Q4_K_M (3.8B) - 21/48
Phi-3.1-mini-4k-instruct-Q4_K_S (3.8B) - 20/48
Phi-3.1-mini-4k-instruct-Q5_K_L (3.8B) - 19/48
Phi-3.1-mini-4k-instruct-Q5_K_M (3.8B) - 20/48
Phi-3.1-mini-4k-instruct-Q5_K_S (3.8B) - 21/48
Phi-3.1-mini-4k-instruct-Q6_K (3.8B) - 20/48
Phi-3.1-mini-4k-instruct-Q6_K_L (3.8B) - 20/48
Phi-3.1-mini-4k-instruct-Q8_0 (3.8B) - 19/48
Phi-3.1-mini-128k-instruct-IQ2_S (3.8B) - 1/48
Phi-3.1-mini-128k-instruct-IQ2_XS (3.8B) - 1/48
Phi-3.1-mini-128k-instruct-IQ3_XXS (3.8B) - 19/48
Phi-3-medium-128k-instruct-f32 (14B) - 30/48
Phi-3-medium-128k-instruct-IQ1_M (14B) - 3/48
Phi-3-medium-128k-instruct-IQ1_S (14B) - 5/48
Phi-3-medium-128k-instruct-IQ2_M (14B) - 22/48
Phi-3-medium-128k-instruct-IQ2_S (14B) - 22/48
Phi-3-medium-128k-instruct-IQ2_XS (14B) - 20/48
Phi-3-medium-128k-instruct-IQ2_XXS (14B) - 11/48
Phi-3-medium-128k-instruct-IQ3_M (14B) - 31/48
Phi-3-medium-128k-instruct-IQ3_S (14B) - 30/48
Phi-3-medium-128k-instruct-IQ3_XS (14B) - 29/48
Phi-3-medium-128k-instruct-IQ3_XXS (14B) - 30/48
Phi-3-medium-128k-instruct-IQ4_NL (14B) - 31/48
Phi-3-medium-128k-instruct-IQ4_XS (14B) - 31/48
Phi-3-medium-128k-instruct-Q2_K (14B) - 23/48
Phi-3-medium-128k-instruct-Q3_K_L (14B) - 30/48
Phi-3-medium-128k-instruct-Q3_K_M (14B) - 29/48
Phi-3-medium-128k-instruct-Q3_K_S (14B) - 28/48
Phi-3-medium-128k-instruct-Q4_K_M (14B) - 31/48
Phi-3-medium-128k-instruct-Q4_K_S (14B) - 32/48
Phi-3-medium-128k-instruct-Q5_K_M (14B) - 31/48
Phi-3-medium-128k-instruct-Q5_K_S (14B) - 32/48
Phi-3-medium-128k-instruct-Q6_K (14B) - 32/48
Phi-3-medium-128k-instruct-Q8_0 (14B) - 32/48
Phi-3-medium-4k-instruct-f32 (14B) - 29/48
Phi-3-medium-4k-instruct-IQ1_M (14B) - 2/48
Phi-3-medium-4k-instruct-IQ1_S (14B) - 1/48
Phi-3-medium-4k-instruct-IQ2_M (14B) - 22/48
Phi-3-medium-4k-instruct-IQ2_S (14B) - 25/48
Phi-3-medium-4k-instruct-IQ2_XS (14B) - 21/48
Phi-3-medium-4k-instruct-IQ2_XXS (14B) - 13/48
Phi-3-medium-4k-instruct-IQ3_M (14B) - 26/48
Phi-3-medium-4k-instruct-IQ3_S (14B) - 27/48
Phi-3-medium-4k-instruct-IQ3_XS (14B) - 27/48
Phi-3-medium-4k-instruct-IQ3_XXS (14B) - 31/48
Phi-3-medium-4k-instruct-IQ4_NL (14B) - 27/48
Phi-3-medium-4k-instruct-IQ4_XS (14B) - 26/48
Phi-3-medium-4k-instruct-Q2_K (14B) - 22/48
Phi-3-medium-4k-instruct-Q3_K_L (14B) - 27/48
Phi-3-medium-4k-instruct-Q3_K_M (14B) - 28/48
Phi-3-medium-4k-instruct-Q3_K_S (14B) - 23/48
Phi-3-medium-4k-instruct-Q4_K_M (14B) - 29/48
Phi-3-medium-4k-instruct-Q4_K_S (14B) - 29/48
Phi-3-medium-4k-instruct-Q5_K_M (14B) - 29/48
Phi-3-medium-4k-instruct-Q5_K_S (14B) - 29/48
Phi-3-medium-4k-instruct-Q6_K (14B) - 29/48
Phi-3-medium-4k-instruct-Q8_0 (14B) - 29/48
Phi-3-mini-4k-instruct-v0.3-f32 (3.8B) - 12/48
Phi-3-mini-4k-instruct-v0.3-IQ2_M (3.8B) - 0/48
Phi-3-mini-4k-instruct-v0.3-IQ2_S (3.8B) - 2/48
Phi-3-mini-4k-instruct-v0.3-IQ2_XS (3.8B) - 1/48
Phi-3-mini-4k-instruct-v0.3-IQ3_M (3.8B) - 11/48
Phi-3-mini-4k-instruct-v0.3-IQ3_XS (3.8B) - 7/48
Phi-3-mini-4k-instruct-v0.3-IQ3_XXS (3.8B) - 17/48
Phi-3-mini-4k-instruct-v0.3-IQ4_XS (3.8B) - 11/48
Phi-3-mini-4k-instruct-v0.3-Q2_K (3.8B) - 8/48
Phi-3-mini-4k-instruct-v0.3-Q3_K_L (3.8B) - 15/48
Phi-3-mini-4k-instruct-v0.3-Q3_K_M (3.8B) - 13/48
Phi-3-mini-4k-instruct-v0.3-Q3_K_S (3.8B) - 17/48
Phi-3-mini-4k-instruct-v0.3-Q4_K_M (3.8B) - 15/48
Phi-3-mini-4k-instruct-v0.3-Q4_K_S (3.8B) - 16/48
Phi-3-mini-4k-instruct-v0.3-Q5_K_M (3.8B) - 15/48
Phi-3-mini-4k-instruct-v0.3-Q5_K_S (3.8B) - 17/48
Phi-3-mini-4k-instruct-v0.3-Q6_K (3.8B) - 13/48
Phi-3-mini-4k-instruct-v0.3-Q8_0 (3.8B) - 12/48
Phi-3-mini-4k-instruct-fp32 (3.8B) - 24/48
Phi-3-mini-4k-instruct-IQ1_M (3.8B) - 0/48
Phi-3-mini-4k-instruct-IQ1_S (3.8B) - 0/48
Phi-3-mini-4k-instruct-IQ2_M (3.8B) - 4/48
Phi-3-mini-4k-instruct-IQ2_S (3.8B) - 6/48
Phi-3-mini-4k-instruct-IQ2_XS (3.8B) - 1/48
Phi-3-mini-4k-instruct-IQ2_XXS (3.8B) - 3/48
Phi-3-mini-4k-instruct-IQ3_M (3.8B) - 19/48
Phi-3-mini-4k-instruct-IQ3_S (3.8B) - 18/48
Phi-3-mini-4k-instruct-IQ3_XS (3.8B) - 15/48
Phi-3-mini-4k-instruct-IQ3_XXS (3.8B) - 21/48
Phi-3-mini-4k-instruct-IQ4_NL (3.8B) - 24/48
Phi-3-mini-4k-instruct-IQ4_XS (3.8B) - 24/48
Phi-3-mini-4k-instruct-Q2_K (3.8B) - 11/48
Phi-3-mini-4k-instruct-Q3_K_L (3.8B) - 20/48
Phi-3-mini-4k-instruct-Q3_K_M (3.8B) - 20/48
Phi-3-mini-4k-instruct-Q3_K_S (3.8B) - 18/48
Phi-3-mini-4k-instruct-Q4_K_M (3.8B) - 23/48
Phi-3-mini-4k-instruct-Q4_K_S (3.8B) - 22/48
Phi-3-mini-4k-instruct-Q5_K_M (3.8B) - 25/48
Phi-3-mini-4k-instruct-Q5_K_S (3.8B) - 26/48
Phi-3-mini-4k-instruct-Q6_K (3.8B) - 24/48
Phi-3-mini-4k-instruct-Q8_0 (3.8B) - 24/48
Phi-3-mini-4k-instruct-old-fp16 (3.8B) - 24/48
Phi-3-mini-4k-instruct-old-IQ1_M (3.8B) - 0/48
Phi-3-mini-4k-instruct-old-IQ1_S (3.8B) - 0/48
Phi-3-mini-4k-instruct-old-IQ2_M (3.8B) - 3/48
Phi-3-mini-4k-instruct-old-IQ2_S (3.8B) - 5/48
Phi-3-mini-4k-instruct-old-IQ2_XS (3.8B) - 1/48
Phi-3-mini-4k-instruct-old-IQ2_XXS (3.8B) - 3/48
Phi-3-mini-4k-instruct-old-IQ3_M (3.8B) - 19/48
Phi-3-mini-4k-instruct-old-IQ3_S (3.8B) - 18/48
Phi-3-mini-4k-instruct-old-IQ3_XS (3.8B) - 15/48
Phi-3-mini-4k-instruct-old-IQ3_XXS (3.8B) - 22/48
Phi-3-mini-4k-instruct-old-IQ4_NL (3.8B) - 24/48
Phi-3-mini-4k-instruct-old-IQ4_XS (3.8B) - 23/48
Phi-3-mini-4k-instruct-old-Q2_K (3.8B) - 9/48
Phi-3-mini-4k-instruct-old-Q3_K_L (3.8B) - 21/48
Phi-3-mini-4k-instruct-old-Q3_K_M (3.8B) - 19/48
Phi-3-mini-4k-instruct-old-Q3_K_S (3.8B) - 18/48
Phi-3-mini-4k-instruct-old-Q4_K_M (3.8B) - 22/48
Phi-3-mini-4k-instruct-old-Q4_K_S (3.8B) - 21/48
Phi-3-mini-4k-instruct-old-Q5_K_M (3.8B) - 25/48
Phi-3-mini-4k-instruct-old-Q5_K_S (3.8B) - 24/48
Phi-3-mini-4k-instruct-old-Q6_K (3.8B) - 24/48
Phi-3-mini-4k-instruct-old-Q8_0 (3.8B) - 24/48
gemma-2-27b-it-IQ2_M (27B) - 28/48
gemma-2-27b-it-IQ2_S (27B) - 25/48
gemma-2-27b-it-IQ2_XS (27B) - 26/48
gemma-2-27b-it-IQ3_M (27B) - 31/48
gemma-2-27b-it-IQ3_XS (27B) - 29/48
gemma-2-27b-it-IQ3_XXS (27B) - 26/48
gemma-2-27b-it-IQ4_XS (27B) - 32/48
gemma-2-27b-it-Q2_K (27B) - 26/48
gemma-2-27b-it-Q2_K_L (27B) - 26/48
gemma-2-27b-it-Q3_K_L (27B) - 31/48
gemma-2-27b-it-Q3_K_M (27B) - 30/48
gemma-2-27b-it-Q3_K_S (27B) - 31/48
gemma-2-27b-it-Q3_K_XL (27B) - 32/48
gemma-2-27b-it-Q4_K_L (27B) - 32/48
gemma-2-27b-it-Q4_K_M (27B) - 33/48
gemma-2-27b-it-Q4_K_S (27B) - 33/48
gemma-2-27b-it-Q5_K_L (27B) - 31/48
gemma-2-27b-it-Q5_K_M (27B) - 33/48
gemma-2-27b-it-Q5_K_S (27B) - 33/48
gemma-2-27b-it-Q6_K (27B) - 33/48
gemma-2-27b-it-Q6_K_L (27B) - 33/48
gemma-2-27b-it-Q8_0 (27B) - 33/48

2024/08/01

google_gemma-2-2b-it (2B) - 18/48
Meta-Llama-3.1-405B-Instruct.i1-IQ1_M (405B) - 25/48
Meta-Llama-3.1-405B-Instruct.i1-IQ1_S (405B) - 17/48
Meta-Llama-3.1-405B-Instruct.i1-IQ2_XXS (405B) - 33/48

2024/07/29

Meta-Llama-3.1-70B-Instruct-IQ1_M (70B) - 11/48
Meta-Llama-3.1-70B-Instruct-IQ2_M (70B) - 30/48
Meta-Llama-3.1-70B-Instruct-IQ2_S (70B) - 27/48
Meta-Llama-3.1-70B-Instruct-IQ2_XS (70B) - 24/48
Meta-Llama-3.1-70B-Instruct-IQ2_XXS (70B) - 18/48
Meta-Llama-3.1-70B-Instruct-IQ3_M (70B) - 34/48
Meta-Llama-3.1-70B-Instruct-IQ3_XS (70B) - 35/48
Meta-Llama-3.1-70B-Instruct-IQ4_XS (70B) - 35/48
Meta-Llama-3.1-70B-Instruct-Q2_K (70B) - 34/48
Meta-Llama-3.1-70B-Instruct-Q2_K_L (70B) - 34/48
Meta-Llama-3.1-70B-Instruct-Q3_K_L (70B) - 37/48
Meta-Llama-3.1-70B-Instruct-Q3_K_M (70B) - 37/48
Meta-Llama-3.1-70B-Instruct-Q3_K_S (70B) - 36/48
Meta-Llama-3.1-70B-Instruct-Q3_K_XL (70B) - 37/48
Meta-Llama-3.1-70B-Instruct-Q4_K_L (70B) - 37/48
Meta-Llama-3.1-70B-Instruct-Q4_K_M (70B) - 38/48
Meta-Llama-3.1-70B-Instruct-Q4_K_S (70B) - 36/48
Meta-Llama-3.1-70B-Instruct-Q5_K_L (70B) - 34/48
Meta-Llama-3.1-70B-Instruct-Q5_K_M (70B) - 34/48
Meta-Llama-3.1-70B-Instruct-Q5_K_S (70B) - 35/48
Meta-Llama-3.1-70B-Instruct-Q6_K (70B) - 35/48
Meta-Llama-3.1-70B-Instruct-Q6_K_L (70B) - 35/48
Meta-Llama-3.1-8B-Instruct-f32 (8B) - 19/48
Meta-Llama-3.1-8B-Instruct-IQ2_M (8B) - 11/48
Meta-Llama-3.1-8B-Instruct-IQ3_M (8B) - 14/48
Meta-Llama-3.1-8B-Instruct-IQ3_XS (8B) - 14/48
Meta-Llama-3.1-8B-Instruct-IQ4_XS (8B) - 17/48
Meta-Llama-3.1-8B-Instruct-Q2_K (8B) - 7/48
Meta-Llama-3.1-8B-Instruct-Q2_K_L (8B) - 9/48
Meta-Llama-3.1-8B-Instruct-Q3_K_L (8B) - 17/48
Meta-Llama-3.1-8B-Instruct-Q3_K_M (8B) - 16/48
Meta-Llama-3.1-8B-Instruct-Q3_K_S (8B) - 15/48
Meta-Llama-3.1-8B-Instruct-Q3_K_XL (8B) - 16/48
Meta-Llama-3.1-8B-Instruct-Q4_K_L (8B) - 20/48
Meta-Llama-3.1-8B-Instruct-Q4_K_M (8B) - 20/48
Meta-Llama-3.1-8B-Instruct-Q4_K_S (8B) - 21/48
Meta-Llama-3.1-8B-Instruct-Q5_K_L (8B) - 20/48
Meta-Llama-3.1-8B-Instruct-Q5_K_M (8B) - 21/48
Meta-Llama-3.1-8B-Instruct-Q5_K_S (8B) - 17/48
Meta-Llama-3.1-8B-Instruct-Q6_K (8B) - 21/48
Meta-Llama-3.1-8B-Instruct-Q6_K_L (8B) - 21/48
Meta-Llama-3.1-8B-Instruct-Q8_0 (8B) - 18/48
turboderp_Mistral-Large-Instruct-2407-123B-exl2_4.25bpw (123B) - 35/48

2024/07/27

DeepSeek-V2-Chat-0628-IQ2_XS (236B) - 17/48
Mistral-Large-Instruct-2407-IQ1_M (123B) - 26/48
Mistral-Large-Instruct-2407-IQ2_XS (123B) - 33/48
Mistral-Large-Instruct-2407-IQ2_XXS (123B) - 29/48
Mistral-Large-Instruct-2407.i1-IQ4_XS (123B) - 37/48
Mistral-Large-Instruct-2407.i1-Q6_K (123B) - 37/48

2024/07/25

Mistral-Large-Instruct-2407.i1-Q4_K_S (123B) - 38/48
mistralai_mathstral-7B-v0.1 (7B) - 12/48

2024/07/24

turboderp_Llama-3.1-70B-Instruct-exl2_2.5bpw (70B) - 29/48
turboderp_Llama-3.1-70B-Instruct-exl2_4.0bpw (70B) - 32/48
turboderp_Llama-3.1-70B-Instruct-exl2_4.5bpw (70B) - 35/48
turboderp_Llama-3.1-70B-Instruct-exl2_6.0bpw (70B) - 36/48

2024/07/23

hugging-quants_Meta-Llama-3.1-70B-Instruct-AWQ-INT4 (70B) - 35/48
meta-llama_Meta-Llama-3.1-8B (8B) - 6/48
meta-llama_Meta-Llama-3.1-8B-Instruct (8B) - 24/48

2024/07/22

Mistral-Nemo-Instruct-2407-Q8_0 (12B) - 23/48

2024/07/20

Athene-70B.i1-Q4_K_M (70B) - 32/48

2024/07/18

mistralai_Mistral-Nemo-Instruct-2407 (12B) - 24/48

2024/07/15

DeepSeek-Coder-V2-Instruct-IQ2_XS (236B) - 19/48
h2oai_h2o-danube3-4b-chat (4B) - 7/48
h2oai_h2o-danube3-500m-chat (0.5B) - 0/48

2024/07/11

turboderp_gemma-2-27b-it-exl2_3.0bpw (27B) - 26/48
turboderp_gemma-2-27b-it-exl2_8.0bpw (27B) - 32/48

2024/07/03

internlm_internlm2_5-7b-chat (7B) - 26/48
microsoft_Phi-3-mini-128k-instruct-20240701 (3.8B) - 16/48
microsoft_Phi-3-mini-4k-instruct-20240701 (3.8B) - 20/48

2024/07/02

UCLA-AGI_Gemma-2-9B-It-SPPO-Iter3 (9B) - 29/48

2024/07/01

c4ai-command-r-plus.i1-IQ4_XS (104B) - 29/48
google_gemma-2-27b-it (27B) - 33/48
google_gemma-2-27b-it (27B) - 30/48
google_gemma-2-9b-it (9B) - 29/48
google_gemma-2-9b-it (9B) - 27/48
google_gemma-2-9b-it (9B) - 26/48

2024/06/28

gustavecortal_oneirogen-7B (7B) - 21/48
magnum-72b-v1-Q4_K_M (72B) - 33/48

2024/06/26

Meta-Llama-3-70B-Instruct-abliterated-v3.5.i1-Q4_K_S (70B) - 30/48
UCLA-AGI_Llama-3-Instruct-8B-SPPO-Iter3 (8B) - 19/48
ZeusLabs_L3-Aethora-15B-V2 (15B) - 17/48

2024/06/25

dolphin-2.9.3-Yi-1.5-34B-32k-Q8_0 (34B) - 33/48
internlm_internlm2-wqx-20b (20B) - 13/48
Llama3-70B-ShiningValiant2.i1-Q4_K_M (70B) - 30/48
mlfoundations_tabula-8b (8B) - 11/48
nyunai_nyun-c2-llama3-56B (56B) - 11/48
Qwen2-72B-Instruct-Q4_K_M (72B) - 32/48
Qwen_Qwen2-0.5B-Instruct (0.5B) - 0/48
Qwen_Qwen2-1.5B-Instruct (1.5B) - 9/48
Qwen_Qwen2-7B-Instruct (7B) - 27/48
turboderp_llama3-turbcat-instruct-8b (8B) - 18/48

2024/05/24

openchat_openchat-3.6-8b-20240522 (8B) - 17/48

2024/05/23

CohereForAI_aya-23-35B (35B) - 20/48
kubernetes-bad_Mistral-7B-Instruct-v0.3 (7B) - 17/48

2024/05/21

microsoft_Phi-3-medium-128k-instruct (14B) - 31/48
microsoft_Phi-3-medium-4k-instruct (14B) - 29/48
microsoft_Phi-3-small-128k-instruct (7B) - 21/48
microsoft_Phi-3-small-8k-instruct (7B) - 26/48
microsoft_Phi-3-vision-128k-instruct (4.2B) - 21/48

2024/05/20

01-ai_Yi-1.5-34B-Chat-16K (34B) - 33/48

2024/05/19

Salesforce_SFR-Iterative-DPO-LLaMA-3-8B-R (8B) - 21/48
Smaug-Llama-3-70B-Instruct.i1-Q4_K_S (70B) - 33/48
tiiuae_falcon-11B (11B) - 7/48

2024/05/12

01-ai_Yi-1.5-34B-Chat (34B) - 33/48
01-ai_Yi-1.5-6B-Chat (6B) - 21/48
01-ai_Yi-1.5-9B-Chat (9B) - 27/48

2024/05/10

Dracones_Llama-3-Lumimaid-70B-v0.1_exl2_4.5bpw (70B) - 29/48

2024/05/07

bartowski_CodeQwen1.5-7B-Chat-exl2_8_0 (7B) - 1/48
Llama3-TenyxChat-70B.i1-Q4_K_S (70B) - 33/48
LoneStriker_Llama-3-70B-Instruct-Gradient-262k-6.0bpw-h6-exl2 (70B) - 24/48
LoneStriker_Llama-3-70B-Instruct-Gradient-524k-6.0bpw-h6-exl2 (70B) - 25/48
Qwen_CodeQwen1.5-7B-Chat (7B) - 1/48
turboderp_command-r-plus-103B-exl2_2.5bpw (104B) - 22/48
turboderp_command-r-plus-103B-exl2_3.0bpw (104B) - 29/48

2024/05/06

failspy_kappa-3-phi-abliterated (3.8B) - 18/48
Llama-3-Giraffe-70B-Instruct.i1-Q4_K_S (70B) - 32/48
turboderp_Cat-Llama-3-70B-instruct-exl2_5.0bpw (70B) - 33/48
turboderp_command-r-plus-103B-exl2_3.5bpw (104B) - 28/48
turboderp_command-r-plus-103B-exl2_4.5bpw (104B) - 28/48

2024/05/05

LoneStriker_dolphin-2.2-yi-34b-200k-8.0bpw-h8-exl2 (34B) - 26/48
LoneStriker_Yi-34B-200K-8.0bpw-h8-exl2 (34B) - 26/48
LoneStriker_Yi-34B-Chat-8.0bpw-h8-exl2 (34B) - 29/48
Meta-Llama-3-120B-Instruct.Q4_K_M (120B) - 31/48
microsoft_Phi-3-mini-128k-instruct (3.8B) - 18/48
microsoft_Phi-3-mini-128k-instruct (3.8B) - 17/48
Platypus2-70B.i1-Q4_K_M (70B) - 27/48

2024/05/04

gradientai_Llama-3-8B-Instruct-Gradient-1048k (8B) - 8/48
internlm_internlm2-chat-1_8b (1.8B) - 4/48
internlm_internlm2-chat-1_8b-sft (1.8B) - 5/48
internlm_internlm2-chat-20b (20B) - 21/48
internlm_internlm2-chat-20b-sft (20B) - 19/48
internlm_internlm2-chat-7b (7B) - 19/48
internlm_internlm2-chat-7b-sft (7B) - 17/48
ISTA-DASLab_c4ai-command-r-plus-AQLM-2Bit-1x16 (104B) - 27/48
ISTA-DASLab_c4ai-command-r-v01-AQLM-2Bit-1x16 (35B) - 10/48
ISTA-DASLab_Meta-Llama-3-70B-Instruct-AQLM-2Bit-1x16 (70B) - 33/48
ISTA-DASLab_Meta-Llama-3-8B-Instruct-AQLM-2Bit-1x16 (8B) - 12/48
jieliu_Storm-7B (7B) - 18/48
miquliz-120b-v2.0.Q4_K_M (120B) - 30/48
TheProfessor-155b.i1-IQ3_XS (155B) - 18/48

2024/05/03

ai21labs_Jamba-v0.1 (52B) - 19/48
gradientai_Llama-3-8B-Instruct-Gradient-1048k (8B) - 13/48
hjhj3168_Llama-3-8b-Orthogonalized-exl2 (8B) - 15/48
Llama3-ChatQA-1.5-70B.Q4_K_M (70B) - 29/48
Llama3-ChatQA-1.5-70B.Q4_K_M (70B) - 27/48
LoneStriker_deepseek-coder-33b-instruct-6.0bpw-h6-exl2 (33B) - 5/48
NousResearch_Hermes-2-Pro-Llama-3-8B (8B) - 15/48
nvidia_ChatQA-1.5-8B (8B) - 14/48
nvidia_ChatQA-1.5-8B (8B) - 13/48
TheBloke_deepseek-coder-33B-instruct-AWQ (33B) - 4/48
turboderp_Phi-3-mini-128k-instruct-exl2_2.5bpw (3.8B) - 1/48
turboderp_Phi-3-mini-128k-instruct-exl2_3.0bpw (3.8B) - 4/48
turboderp_Phi-3-mini-128k-instruct-exl2_4.0bpw (3.8B) - 14/48
turboderp_Phi-3-mini-128k-instruct-exl2_5.0bpw (3.8B) - 17/48
turboderp_Phi-3-mini-128k-instruct-exl2_6.0bpw (3.8B) - 17/48

2024/04/28

EleutherAI_gpt-j-6b (6B) - 0/48
EleutherAI_gpt-neo-1.3B (1.3B) - 0/48
EleutherAI_gpt-neo-2.7B (2.7B) - 0/48
EleutherAI_gpt-neox-20b (20B) - 0/48
facebook_galactica-1.3b (1.3B) - 0/48
facebook_opt-13b (13B) - 0/48
facebook_opt-30b (30B) - 0/48
facebook_opt-6.7b (6.7B) - 0/48
GeorgiaTechResearchInstitute_galactica-30b-evol-instruct-70k (30B) - 13/48
GeorgiaTechResearchInstitute_galactica-30b-evol-instruct-70k (30B) - 7/48
openai-community_gpt2 (0.124B) - 0/48
openai-community_gpt2-large (0.774B) - 0/48
openai-community_gpt2-medium (0.355B) - 0/48
openai-community_gpt2-xl (1.5B) - 0/48

2024/04/27

BAAI_Bunny-Llama-3-8B-V (8B) - 20/48
LoneStriker_dolphin-2.9-llama3-70b-6.0bpw-h6-exl2 (70B) - 33/48
LoneStriker_OpenBioLLM-Llama3-70B-6.0bpw-h6-exl2 (70B) - 34/48
mzbac_llama-3-8B-Instruct-function-calling (8B) - 17/48
Undi95_Meta-Llama-3-70B-Instruct-hf (70B) - 33/48

2024/04/26

Meta-Llama-3-70B-Instruct-IQ2_M (70B) - 32/48
Meta-Llama-3-70B-Instruct-IQ2_S (70B) - 32/48
Meta-Llama-3-70B-Instruct-IQ2_XS (70B) - 31/48
Meta-Llama-3-70B-Instruct-IQ3_M (70B) - 31/48
Meta-Llama-3-70B-Instruct-IQ3_S (70B) - 32/48
Meta-Llama-3-70B-Instruct-IQ3_XS (70B) - 33/48
Meta-Llama-3-70B-Instruct-IQ3_XXS (70B) - 33/48
Meta-Llama-3-70B-Instruct-IQ4_NL (70B) - 32/48
Meta-Llama-3-70B-Instruct-IQ4_XS (70B) - 32/48
Meta-Llama-3-70B-Instruct-Q2_K (70B) - 31/48
Meta-Llama-3-70B-Instruct-Q3_K_L (70B) - 32/48
Meta-Llama-3-70B-Instruct-Q3_K_M (70B) - 32/48
Meta-Llama-3-70B-Instruct-Q3_K_S (70B) - 33/48
Meta-Llama-3-70B-Instruct-Q4_K_S (70B) - 34/48
Qwen1.5-110B-Chat-Q4_K_M (110B) - 29/48
Weyaxi_Einstein-v6.1-Llama3-8B (8B) - 20/48

2024/04/25

34b-beta.Q8_0 (34B) - 29/48
bhenrym14_airoboros-3_1-yi-34b-200k (34B) - 23/48
CausalLM-RP-34B.q8_0 (34B) - 26/48
cloudyu_Phoenix_DPO_60B (60B) - 31/48
ggml-alpaca-dragon-72b-v1-q4_k_m (72B) - 17/48
gradientai_Llama-3-8B-Instruct-262k (8B) - 13/48
MoMo-72B-lora-1.8.6-DPO-Q4_K_M (72B) - 22/48
platypus-yi-34b.Q8_0 (34B) - 34/48
TheBloke_Helion-4x34B-GPTQ (4x34B) - 27/48

2024/04/24

Ein-72B-v0.1-full.Q4_K_M (72B) - 20/48
falcon-180b-chat.Q4_K_M (180B) - 30/48
falcon-180b.Q4_K_M (180B) - 21/48
grok-1-IQ2_XS (314B) - 17/48
ISTA-DASLab_Meta-Llama-3-8B-Instruct-AQLM-2Bit-1x16 (8B) - 8/48
lightblue_suzume-llama-3-8B-multilingual (8B) - 19/48
LoneStriker_Smaug-72B-v0.1-6.0bpw-h6-exl2 (72B) - 28/48
MultiVerse_70B.Q4_K_M (70B) - 24/48
Rhea-72b-v0.5-Q4_K_M (72B) - 30/48
tiiuae_falcon-40b-instruct (40B) - 6/48
xwin-lm-70b-v0.1.Q4_K_M (70B) - 23/48
zephyr-orpo-141b-A35b-v0.1.Q4_K_M (141B) - 19/48

2024/04/23

meraGPT_mera-mix-4x7B (4x7B) - 22/48
microsoft_Phi-3-mini-128k-instruct (3.8B) - 19/48
microsoft_Phi-3-mini-4k-instruct (3.8B) - 23/48
Orenguteng_Lexi-Llama-3-8B-Uncensored (8B) - 18/48

About

This test consists of 48 manually written multiple-choice questions. It evaluates a combination of academic knowledge and logical reasoning.

Compared to MMLU, it has the advantage of not being in any training dataset, and the disadvantage of being much smaller. Compared to lmsys chatbot arena, it is harsher on small models like Starling-LM-7B-beta that write nicely formatted replies but don't have much knowledge.

The correct Jinja2 instruction template is used for each model, as autodetected by text-generation-webui from the model's metadata. For base models without a template, Alpaca is used. The questions are evaluated using the /v1/internal/logits endpoint in the project's API.

The questions are private.

Limitations

This benchmark does not evaluate code generation, non-English languages, role-playing, RAG, and long context understanding. The performance in those areas may have a weak or nonexistent correlation with what is being measured.

Do you find this useful? Consider making a donation.