Datasets:

mteb-pt
/

mteb-pt-results

rank int64 1 93	model large_stringlengths 5 61	params large_stringlengths 3 5	type large_stringclasses 2 values	HateBR float64 0.55 0.88	FactckBrClassification float64 0.35 0.61	ToxSynPT float64 0.59 0.91	PortuLexRRIP float64 0.29 0.56	BrighterEmotionMultilabelClassification float64 0.18 0.4	AssinRTE float64 0.37 0.89	InferBR float64 0.39 0.91	AssinSTS float64 0.25 0.82	Assin2STS float64 0.26 0.83	MedPTClustering float64 0.54 0.89	WikipediaPTCategoriesClusteringP2P float64 0.36 0.8	JurisTCUClusteringP2P float64 0.14 0.46	SciELOClusteringP2P float64 0.13 0.83	StackoverflowPtClustering float64 0.29 0.65	MedPTRetrieval float64 0.02 0.89	FaQuADIR float64 0.06 0.87	Quati float64 0.01 0.69	FaqBacenRetrieval float64 0.02 0.83	JurisTCU float64 0.01 0.66	BRTaxQAR float64 0.01 0.45	QuatiReranking float64 0.19 0.69	JurisTCUReranking float64 0.14 0.6	mean_22 float64 0.25 0.68
1	gemini-embedding-001	---	C	0.8761	0.5645	0.8903	0.4505	0.3996	0.8294	0.8393	0.8184	0.8232	0.7917	0.6784	0.3212	0.8064	0.5878	0.877	0.8561	0.6554	0.7532	0.6096	0.3662	0.6552	0.5551	0.682
2	Qwen3-Embedding-8B	7.6B	O	0.8378	0.5409	0.8744	0.3453	0.2767	0.8664	0.9079	0.7869	0.8267	0.7205	0.7992	0.3682	0.7776	0.5775	0.8248	0.8379	0.6413	0.7053	0.6206	0.4241	0.6298	0.5596	0.6704
3	KaLM-Embedding-Gemma3-12B-2511	11.8B	O	0.856	0.5722	0.9008	0.3969	0.2828	0.8862	0.8986	0.8065	0.8178	0.733	0.7722	0.3696	0.8263	0.5239	0.8741	0.8311	0.5596	0.7297	0.583	0.3638	0.6215	0.5368	0.6701
4	voyage-context-4	---	C	0.8614	0.5653	0.9013	0.4733	0.2981	0.8406	0.8112	0.7861	0.8016	0.6604	0.6053	0.4286	0.6601	0.4487	0.8919	0.8738	0.6901	0.8062	0.6338	0.386	0.6861	0.5768	0.6676
5	Octen-Embedding-8B	7.6B	O	0.8348	0.5317	0.8754	0.356	0.2771	0.864	0.9098	0.7856	0.8229	0.7229	0.7919	0.3752	0.7776	0.5736	0.8492	0.8229	0.6442	0.7174	0.6007	0.3728	0.6344	0.5426	0.6674
6	Qwen3-Embedding-4B	4.0B	O	0.8192	0.505	0.8297	0.328	0.2865	0.866	0.8939	0.7933	0.8208	0.8863	0.7257	0.4111	0.741	0.6255	0.773	0.8099	0.6101	0.6546	0.6163	0.4195	0.5975	0.5539	0.6621
7	voyage-context-3	---	C	0.8482	0.5714	0.8883	0.4444	0.2872	0.8393	0.7698	0.7772	0.7758	0.6454	0.6545	0.4099	0.6484	0.486	0.8907	0.8527	0.6706	0.7959	0.6244	0.3303	0.6711	0.5754	0.6571
8	voyage-3-large	---	C	0.8484	0.583	0.887	0.4608	0.2714	0.8036	0.7763	0.7498	0.7806	0.6418	0.6351	0.4254	0.6691	0.4337	0.8806	0.8626	0.6678	0.7986	0.6159	0.384	0.67	0.5678	0.6552
9	voyage-4-large	---	C	0.8423	0.5402	0.8927	0.4342	0.2865	0.8003	0.7891	0.7617	0.7676	0.7982	0.5377	0.4555	0.6611	0.542	0.894	0.8584	0.6838	0.7337	0.5473	0.3639	0.6661	0.5153	0.6532
10	SFR-Embedding-Mistral	7.1B	O	0.8162	0.4467	0.8962	0.3569	0.3091	0.8616	0.8446	0.8044	0.7994	0.8069	0.7091	0.307	0.7211	0.6526	0.8175	0.8295	0.626	0.6982	0.5414	0.3495	0.6254	0.5315	0.6523
11	BidirLM-1.7B-Embedding	1.7B	O	0.8229	0.5023	0.8981	0.4461	0.2975	0.8552	0.8595	0.7655	0.7998	0.8573	0.7397	0.3178	0.7811	0.6291	0.8313	0.7583	0.5688	0.673	0.5483	0.2656	0.5859	0.5261	0.6513
12	BOOM_4B_v1	4.0B	O	0.8519	0.5556	0.8528	0.3779	0.2775	0.8555	0.7687	0.754	0.7418	0.7757	0.7329	0.3175	0.7664	0.5463	0.8134	0.8623	0.6066	0.7242	0.6408	0.3267	0.5982	0.5602	0.6503
13	embeddinggemma-300m	308M	O	0.8304	0.5691	0.8593	0.4223	0.3237	0.8757	0.8734	0.7886	0.7986	0.7194	0.6861	0.2942	0.7002	0.545	0.7771	0.8464	0.6074	0.6949	0.6207	0.3748	0.5693	0.5023	0.649
14	codestral-embed	---	C	0.8339	0.5699	0.8793	0.4307	0.2854	0.8414	0.864	0.8061	0.8114	0.6588	0.6194	0.2761	0.6755	0.4493	0.8874	0.8117	0.5937	0.8262	0.6369	0.3268	0.6061	0.5787	0.6486
15	Linq-Embed-Mistral	7.1B	O	0.8267	0.4544	0.9076	0.4172	0.2922	0.8556	0.8507	0.8034	0.7978	0.7012	0.757	0.2332	0.6815	0.5668	0.815	0.8403	0.6037	0.7419	0.6227	0.2723	0.6181	0.581	0.6473
16	text-embedding-3-large	---	C	0.8634	0.5521	0.8976	0.4803	0.2988	0.8271	0.7563	0.792	0.7605	0.7219	0.6136	0.3235	0.7003	0.5093	0.8457	0.7895	0.6357	0.7458	0.6085	0.2687	0.6389	0.5584	0.6449
17	jina-embeddings-v5-text-small	596M	O	0.7969	0.4347	0.7737	0.2874	0.2597	0.8576	0.851	0.7766	0.8161	0.799	0.7577	0.3404	0.7646	0.6167	0.78	0.8204	0.5937	0.6874	0.6022	0.3597	0.6236	0.5579	0.6435
18	voyage-4	---	C	0.8394	0.534	0.8897	0.4069	0.2889	0.811	0.7837	0.7382	0.7999	0.7489	0.55	0.4558	0.6658	0.4906	0.8804	0.8448	0.6279	0.737	0.5436	0.3697	0.6358	0.5107	0.6433
19	multilingual-e5-large-instruct	560M	O	0.826	0.5568	0.8943	0.3212	0.3297	0.8771	0.8274	0.8076	0.8058	0.7231	0.7882	0.3668	0.7214	0.5064	0.7693	0.8115	0.561	0.6922	0.5788	0.182	0.5934	0.5589	0.6409
20	SFR-Embedding-2_R	7.1B	O	0.8151	0.4459	0.8922	0.4246	0.3263	0.8512	0.7326	0.7761	0.7445	0.766	0.7621	0.2376	0.7193	0.6356	0.8021	0.7391	0.5646	0.6942	0.6123	0.3937	0.5627	0.5745	0.6397
21	gte-Qwen2-7B-instruct	7.1B	O	0.8414	0.5629	0.8793	0.3871	0.3105	0.8463	0.8294	0.782	0.7945	0.7293	0.7659	0.3076	0.7602	0.4569	0.8195	0.7788	0.5806	0.6935	0.5412	0.3001	0.5795	0.5157	0.6392
22	harrier-oss-v1-27b	27.0B	O	0.8774	0.6054	0.8858	0.4742	0.2548	0.7936	0.7972	0.7861	0.8349	0.6487	0.6195	0.4471	0.701	0.4865	0.8809	0.7321	0.5627	0.7083	0.5443	0.3309	0.5738	0.5119	0.639
23	BidirLM-1B-Embedding	1000M	O	0.8224	0.493	0.8994	0.4371	0.3186	0.8636	0.8541	0.7825	0.8025	0.7157	0.7588	0.3468	0.7244	0.4779	0.8119	0.7716	0.5549	0.6742	0.5789	0.2423	0.5786	0.5456	0.6388
24	F2LLM-v2-8B	7.6B	O	0.8402	0.5846	0.8398	0.432	0.3075	0.7846	0.7999	0.7738	0.7847	0.6143	0.523	0.4103	0.5872	0.4838	0.8712	0.7654	0.6492	0.8229	0.6015	0.3409	0.6272	0.5645	0.6368
25	voyage-3.5	---	C	0.8283	0.534	0.8666	0.398	0.2513	0.8043	0.736	0.7296	0.7493	0.6625	0.6673	0.3744	0.72	0.4772	0.8335	0.8475	0.6148	0.7511	0.5911	0.3434	0.6328	0.5578	0.635
26	harrier-oss-v1-0.6b	596M	O	0.8009	0.5471	0.8529	0.3282	0.2735	0.8593	0.8476	0.7712	0.7953	0.7544	0.7151	0.3485	0.7033	0.4984	0.7541	0.8261	0.5969	0.6507	0.5444	0.3455	0.613	0.5251	0.6342
27	F2LLM-v2-14B	14.0B	O	0.8609	0.5788	0.8373	0.4445	0.3014	0.7869	0.8064	0.778	0.7896	0.6015	0.4952	0.4139	0.5689	0.4777	0.8863	0.7609	0.6427	0.8189	0.5911	0.3053	0.635	0.5639	0.6339
28	gemini-embedding-2	---	C	0.8495	0.5703	0.8961	0.4942	0.2513	0.7687	0.6912	0.7479	0.727	0.7354	0.6148	0.2945	0.7308	0.4942	0.8754	0.8603	0.6469	0.7752	0.4891	0.2938	0.6429	0.4785	0.6331
29	F2LLM-v2-4B	4.0B	O	0.8414	0.5815	0.8359	0.4273	0.2942	0.7803	0.7764	0.7599	0.7783	0.7481	0.4833	0.4382	0.5492	0.48	0.8628	0.7582	0.6159	0.8144	0.5902	0.3094	0.6135	0.5607	0.6318
30	PwC-Embedding_expr	560M	O	0.8258	0.556	0.8799	0.3416	0.2984	0.8814	0.8524	0.8052	0.8026	0.677	0.7142	0.402	0.631	0.5736	0.7677	0.7959	0.55	0.6696	0.5248	0.2038	0.5798	0.5045	0.629
31	voyage-finance-2	---	C	0.8176	0.5757	0.8621	0.4588	0.2626	0.7718	0.6947	0.7245	0.7159	0.5658	0.6319	0.3677	0.7291	0.443	0.8457	0.8227	0.5734	0.7678	0.5631	0.4499	0.5967	0.5365	0.6262
32	Octen-Embedding-0.6B	596M	O	0.7854	0.5066	0.8194	0.3353	0.2688	0.8247	0.8034	0.7466	0.7683	0.7828	0.7643	0.3752	0.7488	0.4969	0.7144	0.8051	0.568	0.6485	0.5602	0.3538	0.5749	0.5191	0.6259
33	PIXIE-Rune-v1.0	568M	O	0.8441	0.5154	0.8482	0.4209	0.2684	0.8361	0.6977	0.7798	0.7384	0.7626	0.5452	0.3547	0.5611	0.4689	0.7681	0.8677	0.61	0.6858	0.6306	0.3265	0.6131	0.5709	0.6234
34	Qwen3-Embedding-0.6B	596M	O	0.7849	0.5104	0.8412	0.3291	0.268	0.8336	0.8167	0.7398	0.7776	0.7022	0.7737	0.3656	0.7061	0.5463	0.6907	0.7935	0.5615	0.6317	0.5737	0.3696	0.5655	0.5283	0.6232
35	gte-Qwen2-1.5B-instruct	1.5B	O	0.8197	0.5459	0.8854	0.3289	0.2836	0.8037	0.753	0.7396	0.7459	0.8022	0.7142	0.4118	0.6457	0.4433	0.7339	0.8083	0.5488	0.6714	0.5862	0.3106	0.5613	0.5365	0.6218
36	voyage-law-2	---	C	0.8311	0.5386	0.8571	0.3921	0.2689	0.807	0.7473	0.7524	0.7489	0.6321	0.6284	0.3821	0.6595	0.4702	0.8042	0.8274	0.5571	0.7066	0.5348	0.4172	0.5846	0.5	0.6203
37	snowflake-arctic-embed-l-v2.0	568M	O	0.8364	0.5324	0.8344	0.406	0.2616	0.7942	0.6635	0.7429	0.7281	0.74	0.5907	0.3865	0.6129	0.5051	0.7674	0.8255	0.5849	0.6989	0.6622	0.2775	0.5893	0.6006	0.6201
38	voyage-3.5-lite	---	C	0.8137	0.5622	0.8507	0.4018	0.2577	0.7722	0.6503	0.7068	0.7251	0.6596	0.6618	0.4024	0.6949	0.4395	0.7651	0.8354	0.5971	0.7269	0.5932	0.3506	0.608	0.5587	0.6197
39	text-embedding-3-small	---	C	0.8464	0.5398	0.8452	0.4382	0.2737	0.82	0.6853	0.7706	0.7162	0.6799	0.6425	0.3712	0.6458	0.5108	0.7748	0.8017	0.5676	0.6773	0.5803	0.2808	0.5984	0.5421	0.6186
40	bge-m3	568M	O	0.8206	0.4819	0.8564	0.4271	0.2844	0.8224	0.8248	0.7537	0.7736	0.6189	0.5751	0.3598	0.5776	0.4075	0.7468	0.8293	0.5944	0.6717	0.5909	0.3772	0.5949	0.5576	0.6157
41	voyage-4-lite	---	C	0.808	0.5151	0.8667	0.3921	0.2792	0.7967	0.7142	0.7126	0.732	0.65	0.6136	0.4231	0.7291	0.4808	0.7979	0.8329	0.5742	0.7104	0.4947	0.3428	0.592	0.4817	0.6154
42	F2LLM-v2-1.7B	1.7B	O	0.8325	0.5641	0.8226	0.4266	0.2826	0.7764	0.7599	0.7647	0.7465	0.6037	0.5527	0.387	0.5644	0.4382	0.8437	0.7596	0.5829	0.7896	0.5853	0.2916	0.5953	0.5572	0.6149
43	BidirLM-0.6B-Embedding	596M	O	0.7847	0.4865	0.882	0.4095	0.2796	0.8253	0.8204	0.7385	0.7823	0.6964	0.7263	0.3368	0.7256	0.5113	0.7531	0.7559	0.5311	0.6395	0.513	0.2641	0.5674	0.491	0.6146
44	voyage-3	---	C	0.7876	0.5417	0.8442	0.4299	0.242	0.7469	0.6862	0.6846	0.7293	0.6195	0.5833	0.3531	0.6564	0.4886	0.8409	0.8435	0.5701	0.7412	0.5631	0.3934	0.5803	0.5245	0.6114
45	jua-4B-mixed	4.0B	O	0.8126	0.5139	0.7794	0.3534	0.2432	0.7526	0.7129	0.7276	0.7397	0.7455	0.6118	0.4367	0.5735	0.4625	0.8138	0.7947	0.5958	0.6582	0.5795	0.3685	0.6015	0.5303	0.6094
46	F2LLM-0.6B	596M	O	0.7721	0.5031	0.8479	0.3669	0.2788	0.7789	0.7366	0.745	0.7505	0.725	0.7315	0.2797	0.683	0.5656	0.7019	0.7689	0.5569	0.6258	0.5618	0.2725	0.5533	0.5202	0.6057
47	harrier-oss-v1-270m	268M	O	0.8064	0.5278	0.8382	0.3498	0.2588	0.8066	0.7236	0.7671	0.7614	0.7626	0.6966	0.3994	0.6727	0.4274	0.7017	0.7834	0.5469	0.6096	0.4878	0.2982	0.5603	0.4946	0.6037
48	F2LLM-v2-0.6B	596M	O	0.802	0.5429	0.8076	0.4122	0.2693	0.7736	0.7377	0.7475	0.7414	0.7355	0.5027	0.3896	0.5236	0.4669	0.7933	0.7459	0.5615	0.7635	0.5766	0.2404	0.5579	0.5382	0.6014
49	mistral-embed	---	C	0.817	0.5394	0.8442	0.3799	0.2405	0.7884	0.7027	0.7479	0.696	0.6227	0.6481	0.3013	0.7094	0.4646	0.7799	0.8054	0.5591	0.6648	0.5897	0.1887	0.561	0.5248	0.5989
50	titan-embed-text-v2	---	C	0.7804	0.456	0.8452	0.3615	0.2674	0.7908	0.7805	0.7622	0.7455	0.6166	0.5318	0.3198	0.5976	0.4501	0.7276	0.7482	0.5248	0.6516	0.6065	0.3857	0.5302	0.5527	0.5924
51	multilingual-e5-large	560M	O	0.8204	0.5459	0.8557	0.4302	0.2808	0.7874	0.7673	0.7437	0.7832	0.6091	0.5881	0.2497	0.5824	0.4077	0.6828	0.8444	0.5619	0.6509	0.5356	0.1733	0.5939	0.5047	0.5909
52	granite-embedding-311m-multilingual-r2	312M	O	0.7934	0.5058	0.7318	0.3634	0.2592	0.7751	0.6279	0.7246	0.6863	0.7127	0.6221	0.3461	0.624	0.4522	0.7175	0.77	0.5724	0.6813	0.5565	0.3466	0.5763	0.5383	0.5902
53	jua-4B-legal-only	4.0B	O	0.8217	0.5008	0.7801	0.3478	0.2447	0.7781	0.6875	0.7529	0.7364	0.6381	0.5243	0.4147	0.4751	0.4269	0.7962	0.8297	0.5721	0.6351	0.5877	0.3473	0.5534	0.5286	0.59
54	voyage-3-lite	---	C	0.7889	0.5683	0.8138	0.4321	0.2546	0.7499	0.6288	0.6864	0.731	0.5784	0.6124	0.3969	0.6545	0.4428	0.7077	0.8072	0.5206	0.6966	0.4725	0.3397	0.5563	0.4785	0.5872
55	F2LLM-v2-330M	334M	O	0.7867	0.5162	0.7834	0.4229	0.2362	0.7596	0.6543	0.731	0.7102	0.6987	0.4838	0.3684	0.5309	0.4161	0.7517	0.7427	0.5167	0.7319	0.5826	0.2461	0.5319	0.5374	0.5791
56	multilingual-e5-base	278M	O	0.7939	0.517	0.827	0.418	0.2588	0.7955	0.7019	0.7408	0.7445	0.726	0.5482	0.2954	0.534	0.4457	0.619	0.8217	0.5533	0.6387	0.4908	0.1765	0.5456	0.4657	0.5754
57	text-embedding-005	---	C	0.7871	0.5073	0.8309	0.3756	0.2448	0.7233	0.6413	0.6607	0.6889	0.6311	0.6165	0.3462	0.6008	0.5534	0.7254	0.7737	0.5279	0.6087	0.4608	0.2757	0.5447	0.471	0.5725
58	serafim-100m-portuguese-pt-sentence-encoder-ir	100M	O	0.8179	0.4598	0.81	0.4338	0.244	0.83	0.7581	0.7739	0.769	0.6974	0.6244	0.2746	0.5558	0.3911	0.6391	0.7262	0.4499	0.581	0.5538	0.1618	0.4494	0.5254	0.5694
59	serafim-335m-portuguese-pt-sentence-encoder-ir	335M	O	0.8051	0.488	0.8186	0.4324	0.2563	0.7936	0.7273	0.7548	0.7329	0.6959	0.6115	0.2936	0.5686	0.3786	0.6615	0.7246	0.4634	0.5945	0.5436	0.1588	0.4713	0.522	0.5681
60	multilingual-e5-small	118M	O	0.7695	0.5066	0.8034	0.4232	0.2534	0.8006	0.7206	0.7381	0.7427	0.7298	0.5736	0.3026	0.5605	0.4173	0.4942	0.8281	0.4927	0.5872	0.4942	0.1451	0.5	0.4607	0.5611
61	granite-embedding-107m-multilingual	107M	O	0.7626	0.5154	0.7703	0.4014	0.2557	0.8266	0.6059	0.7474	0.6629	0.7496	0.6213	0.2784	0.6296	0.4427	0.5827	0.7423	0.4833	0.5804	0.511	0.1612	0.4938	0.5053	0.5605
62	granite-embedding-97m-multilingual-r2	97M	O	0.7509	0.5274	0.7098	0.343	0.2562	0.7541	0.5791	0.7012	0.6655	0.625	0.594	0.3728	0.5997	0.5142	0.6522	0.7576	0.495	0.6492	0.4618	0.3246	0.512	0.4643	0.5595
63	serafim-900m-portuguese-pt-sentence-encoder-ir	900M	O	0.7983	0.5008	0.8196	0.3979	0.2317	0.769	0.733	0.7388	0.7489	0.6669	0.5441	0.2548	0.5436	0.3898	0.6844	0.7074	0.46	0.623	0.5668	0.1296	0.4644	0.5327	0.5593
64	serafim-335m-portuguese-pt-sentence-encoder	335M	O	0.813	0.4452	0.8778	0.4566	0.2637	0.8717	0.8665	0.7978	0.8323	0.7488	0.5316	0.2709	0.6323	0.421	0.5567	0.6529	0.3693	0.4443	0.3188	0.1084	0.4154	0.3894	0.5493
65	serafim-900m-portuguese-pt-sentence-encoder	900M	O	0.8103	0.4687	0.8885	0.3999	0.2761	0.8723	0.8666	0.7888	0.8267	0.6826	0.5825	0.248	0.5466	0.3828	0.545	0.6214	0.3559	0.4634	0.3156	0.115	0.4098	0.3862	0.5388
66	F2LLM-v2-160M	159M	O	0.7238	0.5143	0.7359	0.3998	0.2314	0.7471	0.565	0.69	0.6626	0.5911	0.5365	0.3226	0.5325	0.4049	0.6631	0.7244	0.476	0.6486	0.5248	0.1502	0.4643	0.4957	0.5366
67	paraphrase-multilingual-mpnet-base-v2	278M	O	0.8034	0.5317	0.841	0.3993	0.2601	0.8049	0.8388	0.7204	0.7437	0.688	0.4617	0.2326	0.4648	0.4305	0.5542	0.6247	0.4024	0.4635	0.3743	0.1017	0.4419	0.441	0.5284
68	serafim-100m-portuguese-pt-sentence-encoder	100M	O	0.8154	0.4544	0.8328	0.4612	0.2642	0.8646	0.8172	0.7871	0.8214	0.581	0.5317	0.2748	0.5662	0.3686	0.4921	0.6521	0.3227	0.4362	0.3409	0.1177	0.3877	0.4015	0.5269
69	Qwen3-Embedding-0.6B-jua-V2	600M	O	0.7566	0.4942	0.6872	0.3056	0.2199	0.6186	0.5097	0.5817	0.5971	0.7677	0.5716	0.3366	0.5729	0.4339	0.6466	0.7425	0.4707	0.4869	0.4621	0.2047	0.4664	0.4466	0.5173
70	F2LLM-v2-80M	80M	O	0.702	0.5046	0.7083	0.3999	0.2279	0.7328	0.5437	0.6505	0.6127	0.5887	0.5011	0.3076	0.4583	0.3995	0.6317	0.6997	0.4345	0.6364	0.4919	0.0948	0.4274	0.4739	0.5104
71	medlink-bi-encoder	110M	O	0.8026	0.4946	0.792	0.4619	0.2362	0.6841	0.5375	0.616	0.6126	0.7493	0.6053	0.344	0.6281	0.4622	0.4478	0.5999	0.3639	0.3386	0.2409	0.1767	0.3977	0.3278	0.4964
72	LaBSE	471M	O	0.8102	0.4811	0.8056	0.4652	0.2571	0.8077	0.577	0.7083	0.648	0.6537	0.5705	0.2622	0.5864	0.4315	0.4389	0.6462	0.2959	0.3744	0.2999	0.0897	0.3449	0.367	0.4964
73	e5-small-v2	33M	O	0.6413	0.5031	0.6722	0.3539	0.2174	0.6989	0.5096	0.6307	0.6246	0.6119	0.5049	0.198	0.4701	0.4582	0.4955	0.689	0.3534	0.4197	0.467	0.1052	0.3373	0.4309	0.4724
74	mxbai-embed-large-v1	335M	O	0.6654	0.4533	0.6927	0.3514	0.2127	0.6768	0.4518	0.5639	0.5521	0.6881	0.5807	0.1718	0.6135	0.4162	0.507	0.6223	0.3551	0.4257	0.3846	0.1159	0.3957	0.4017	0.4681
75	bge-small-en-v1.5	33M	O	0.6527	0.5216	0.6612	0.3417	0.2213	0.6778	0.4669	0.5784	0.5661	0.733	0.5552	0.2019	0.5396	0.4114	0.4773	0.6357	0.3138	0.4343	0.4512	0.1112	0.3145	0.4279	0.4679
76	bert-large-portuguese-cased	335M	O	0.8272	0.5606	0.8581	0.5575	0.2246	0.6657	0.5337	0.6608	0.6192	0.6854	0.6205	0.3401	0.6166	0.336	0.3564	0.4965	0.198	0.2665	0.1936	0.0406	0.3353	0.2913	0.4675
77	gte-small	33M	O	0.6418	0.5042	0.6492	0.3345	0.2275	0.6566	0.4791	0.5886	0.5631	0.6168	0.5773	0.1668	0.5407	0.4699	0.5036	0.6277	0.3327	0.4236	0.4212	0.1486	0.3292	0.3989	0.4637
78	bert-base-portuguese-cased	110M	O	0.819	0.5259	0.854	0.5367	0.2238	0.6901	0.5149	0.6442	0.6145	0.6547	0.5882	0.3128	0.512	0.3574	0.3145	0.5703	0.2474	0.2252	0.1721	0.0816	0.3275	0.2737	0.4573
79	bert-large-portuguese-cased-legal-mlm-sts-v1.0	335M	O	0.7855	0.4494	0.8241	0.428	0.259	0.8345	0.7378	0.766	0.8086	0.5781	0.4755	0.2679	0.167	0.3774	0.4071	0.423	0.0984	0.331	0.32	0.1177	0.2042	0.3655	0.4557
80	e5-mistral-7b-instruct	7.1B	O	0.7098	0.5429	0.8471	0.3407	0.217	0.7804	0.5007	0.7163	0.5149	0.7046	0.3923	0.2089	0.5867	0.401	0.3329	0.605	0.2848	0.4137	0.1307	0.0186	0.3659	0.3362	0.4523
81	llama-embed-nemotron-8b	7.5B	O	0.7764	0.5571	0.8478	0.4399	0.241	0.5629	0.3946	0.5889	0.5072	0.6209	0.5769	0.3629	0.6562	0.3344	0.5445	0.469	0.0496	0.2067	0.1603	0.3146	0.3274	0.2764	0.4462
82	Legal-BERTimbau-sts-large	335M	O	0.7356	0.4239	0.8173	0.3597	0.2232	0.7331	0.6246	0.5927	0.5176	0.6465	0.6323	0.2312	0.5046	0.2926	0.4265	0.5068	0.2164	0.3069	0.2302	0.0606	0.316	0.3645	0.4438
83	bert-large-portuguese-cased-legal-mlm-mkd-nli-sts-v1	335M	O	0.7832	0.4263	0.8715	0.379	0.2896	0.8411	0.8298	0.7627	0.8112	0.603	0.5124	0.1743	0.3356	0.3861	0.363	0.3099	0.0393	0.1819	0.1714	0.0494	0.2496	0.3002	0.4396
84	Ivysaur	23M	O	0.6164	0.4259	0.6511	0.3415	0.2205	0.6667	0.469	0.6023	0.5632	0.6564	0.5415	0.2004	0.44	0.4074	0.3936	0.5993	0.2913	0.3812	0.4069	0.1024	0.2896	0.3831	0.4386
85	albertina-900m-portuguese-ptbr-encoder	900M	O	0.7764	0.51	0.8566	0.478	0.1821	0.6562	0.494	0.6239	0.5648	0.6276	0.5736	0.2578	0.5664	0.3449	0.247	0.5409	0.175	0.2445	0.2035	0.0487	0.2937	0.2902	0.4344
86	all-MiniLM-L12-v2	33M	O	0.6199	0.4683	0.6589	0.3546	0.2075	0.6511	0.4737	0.5333	0.5683	0.7183	0.5447	0.1831	0.4543	0.4572	0.4008	0.5073	0.2912	0.3663	0.3448	0.0642	0.3118	0.3482	0.4331
87	GIST-all-MiniLM-L6-v2	23M	O	0.5975	0.4467	0.6387	0.3129	0.2149	0.6577	0.4685	0.5489	0.5558	0.755	0.5027	0.2033	0.4958	0.4502	0.3883	0.5208	0.2176	0.3215	0.2976	0.1062	0.2888	0.3161	0.423
88	all-mpnet-base-v2	109M	O	0.635	0.4529	0.6501	0.3464	0.2083	0.6196	0.4582	0.5062	0.5631	0.5757	0.5657	0.1914	0.5631	0.4919	0.3864	0.4801	0.2452	0.3151	0.2676	0.0806	0.3124	0.3214	0.4198
89	legal-bert-pt-br	110M	O	0.7127	0.4181	0.7605	0.3851	0.2365	0.7733	0.5831	0.6675	0.6463	0.6712	0.4997	0.1533	0.3147	0.3874	0.2591	0.3472	0.1439	0.2549	0.2264	0.0088	0.2761	0.3044	0.4105
90	all-MiniLM-L6-v2	23M	O	0.5968	0.4421	0.6471	0.3445	0.1969	0.628	0.4743	0.53	0.5644	0.5431	0.5641	0.161	0.4459	0.456	0.367	0.5115	0.2197	0.2811	0.2792	0.0847	0.2903	0.3147	0.4065
91	Legal-BERTimbau-sts-large-ma-v3	335M	O	0.7661	0.4602	0.8533	0.3351	0.2961	0.815	0.8066	0.7232	0.7901	0.6168	0.4007	0.2055	0.1617	0.3895	0.2962	0.1469	0.0263	0.1648	0.1377	0.0748	0.1923	0.2652	0.4056
92	bert-large-portuguese-cased-legal-tsdae-gpl-nli-sts-MetaKD-v0	335M	O	0.7229	0.351	0.8271	0.3113	0.2057	0.5512	0.5858	0.4564	0.4979	0.6252	0.4631	0.2285	0.1479	0.3517	0.1968	0.1805	0.0264	0.2626	0.2549	0.0886	0.2216	0.339	0.3589
93	paraphrase-multilingual-MiniLM-L12-v2	118M	O	0.5525	0.3973	0.5872	0.3377	0.201	0.3715	0.39	0.2487	0.2566	0.6908	0.3557	0.141	0.1323	0.3231	0.0206	0.0583	0.0144	0.0198	0.0121	0.0143	0.1979	0.1382	0.2482

🇧🇷 MTEB-PT — Benchmark Results

Canonical results store for MTEB-PT, a native Brazilian-Portuguese text-embedding benchmark.

93 models · 22 native PT-BR tasks · 7 categories · no machine translation

What is this?

This repository is the canonical, machine-readable results store for MTEB-PT — a benchmark that evaluates text-embedding models on native Brazilian Portuguese (data created or found in Portuguese; machine-translated corpora such as mMARCO-PT are excluded by construction). Every model is evaluated with mteb on datasets pinned to a revision SHA.

The aggregated score_matrix.parquet (93 models × 22 tasks, browsable in the Dataset Viewer above) is the quickest way to see all scores; the raw per-evaluation JSON artifacts live under results/.

📊 Leaderboard (top 15 by 22-task mean)

Rank	Model	Params	Type	mean₂₂
1	`gemini-embedding-001`	---	C	0.6820
2	`Qwen3-Embedding-8B`	7.6B	O	0.6704
3	`KaLM-Embedding-Gemma3-12B-2511`	11.8B	O	0.6701
4	`voyage-context-4`	---	C	0.6676
5	`Octen-Embedding-8B`	7.6B	O	0.6674
6	`Qwen3-Embedding-4B`	4.0B	O	0.6621
7	`voyage-context-3`	---	C	0.6571
8	`voyage-3-large`	---	C	0.6552
9	`voyage-4-large`	---	C	0.6532
10	`SFR-Embedding-Mistral`	7.1B	O	0.6523
11	`BidirLM-1.7B-Embedding`	1.7B	O	0.6513
12	`BOOM_4B_v1`	4.0B	O	0.6503
13	`embeddinggemma-300m`	308M	O	0.6490
14	`codestral-embed`	---	C	0.6486
15	`Linq-Embed-Mistral`	7.1B	O	0.6473

Type: O = open-weight, C = closed/commercial API. Full 93-model × 22-task table: score_matrix.parquet (viewer above). A mteb-pt/baseline-random-encoder chance floor (mean₂₂ = 0.18) is included in the raw results for reference.

🗂️ The 22 tasks (7 categories)

Category	#	Tasks
Classification	4	HateBR, FactckBr, ToxSynPT, PortuLexRRIP
Multilabel classification	1	BrighterEmotion
Pair classification	2	AssinRTE, InferBR
Semantic textual similarity	2	AssinSTS, Assin2STS
Clustering	5	MedPTClustering, WikipediaPTCategories, JurisTCUClustering, SciELOClustering, StackoverflowPt
Retrieval	6	MedPTRetrieval, FaQuADIR, Quati, FaqBacen, JurisTCU, BRTaxQAR
Reranking	2	QuatiReranking, JurisTCUReranking

Domains span legal, medical, tax, scientific, encyclopedic, and social-media Portuguese. Per-task sources, licenses, and citations are documented in the code repository.

📁 Repository layout & how to use

score_matrix.parquet                     # aggregated 93 x 22 matrix (viewer-friendly)
results/
  {org}__{model}/{revision-sha}/
    {Task}.json                          # per-(model, task) mteb result (the canonical score)

There are 2,800+ per-evaluation JSON files plus per-instance .jsonl dumps behind the confidence intervals.

Load the aggregated matrix:

import pandas as pd
from huggingface_hub import hf_hub_download
df = pd.read_parquet(hf_hub_download("mteb-pt/mteb-pt-results", "score_matrix.parquet", repo_type="dataset"))
df.sort_values("mean_22", ascending=False).head()

Pull the full raw store (all JSONs):

from huggingface_hub import snapshot_download
snapshot_download("mteb-pt/mteb-pt-results", repo_type="dataset", local_dir="mteb-pt-results")

🔗 Links

Interactive leaderboard — https://huggingface.co/spaces/mteb-pt/leaderboard
Code & task definitions — https://github.com/tardellirs/mteb-pt
Organization — https://huggingface.co/mteb-pt
This dataset’s DOI — https://doi.org/10.57967/hf/9377
Code archive (Zenodo DOI) — https://doi.org/10.5281/zenodo.21087217 · archives the code repository, not this results dataset

📄 License

This results dataset: CC-BY-4.0.
Individual task datasets: each retains its original source license (see the paper / code repo task table).
Models evaluated: see each model card.

📚 Citation

@misc{mteb-pt-2026,
  title  = {MTEB-PT: A Text Embedding Benchmark for Brazilian Portuguese},
  author = {Stekel, Tardelli R. C.},
  year   = {2026},
  doi    = {10.5281/zenodo.21087217},  % Zenodo archive of the benchmark code
  url    = {https://doi.org/10.5281/zenodo.21087217}
}

If you use a specific task novel to this benchmark, please also cite the original task dataset.

Downloads last month: 3,082

Number of rows:

Total file size:

11.9 MB