Datasets:

GSMA
/

leaderboard

model stringlengths 9 17	rank int64 1 31	average float64 0.17 0.54	benchmarks_completed int64 3 7	teleqna_score float64 0.16 0.76 ⌀	teleqna_stderr float64 0 0 ⌀	teleqna_n float64 10k 10k ⌀	teletables_score float64 0.06 0.33 ⌀	teletables_stderr float64 0.01 0.02 ⌀	teletables_n float64 500 500 ⌀	oranbench_score float64 0.18 0.75 ⌀	oranbench_stderr float64 0.01 0.01 ⌀	oranbench_n float64 1.5k 1.5k ⌀	srsranbench_score float64 0.09 0.83 ⌀	srsranbench_stderr float64 0.01 0.01 ⌀	srsranbench_n float64 1.5k 1.5k ⌀	telemath_score float64 0.02 0.45	telemath_stderr float64 0.01 0.02	telemath_n int64 500 500	telelogs_score float64 0 0.27	telelogs_stderr float64 0 0.01	telelogs_n int64 864 864	three_gpp_score float64 0.1 0.51	three_gpp_stderr float64 0.01 0.01	three_gpp_n int64 2k 2k
qwen2.5-72b	1	0.539701	7	0.764933	0.004235	10,000	0.292	0.020267	500	0.747111	0.011218	1,500	0.794718	0.010397	1,502	0.454667	0.020973	500	0.273148	0.013033	864	0.451333	0.01109	2,000
mistral-small-24b	2	0.51627	7	0.737367	0.004396	10,000	0.297333	0.020397	500	0.717111	0.011608	1,500	0.762983	0.010954	1,502	0.361333	0.020703	500	0.231096	0.01272	864	0.506667	0.011152	2,000
qwen2.5-32b	3	0.506711	7	0.749867	0.004326	10,000	0.285333	0.020149	500	0.732667	0.011418	1,500	0.759432	0.011028	1,502	0.418667	0.02069	500	0.248843	0.0128	864	0.352167	0.01066	2,000
phi-4-14b	4	0.504528	7	0.730833	0.004428	10,000	0.299333	0.019884	500	0.736889	0.011351	1,500	0.833555	0.009589	1,502	0.400667	0.020919	500	0.197917	0.011454	864	0.3325	0.0105	2,000
gemma3-27b	5	0.504326	7	0.713133	0.004521	10,000	0.333333	0.021061	500	0.715778	0.011624	1,500	0.805593	0.010195	1,502	0.406667	0.02048	500	0.160108	0.01082	864	0.395667	0.010896	2,000
qwen2.5-14b	6	0.485376	7	0.7248	0.004463	10,000	0.294	0.020351	500	0.725556	0.011521	1,500	0.776298	0.010756	1,502	0.324	0.019337	500	0.217978	0.011172	864	0.335	0.010551	2,000
mistral-small-22b	7	0.466597	7	0.669767	0.004695	10,000	0.284	0.020143	500	0.693556	0.011895	1,500	0.76964	0.010859	1,502	0.235333	0.017904	500	0.244213	0.013177	864	0.369667	0.01075	2,000
gemma3-12b	8	0.463852	7	0.6879	0.00463	10,000	0.274	0.019921	500	0.695333	0.01188	1,500	0.80071	0.010306	1,502	0.274	0.017468	500	0.185185	0.011009	864	0.329833	0.010468	2,000
falcon3-10b	9	0.458808	7	0.6866	0.004633	10,000	0.276667	0.020004	500	0.636889	0.012393	1,500	0.767865	0.010893	1,502	0.343333	0.020379	500	0.190972	0.011743	864	0.309333	0.010319	2,000
gemma2-27b	10	0.458493	7	0.718	0.004498	10,000	0.272667	0.019913	500	0.707111	0.01175	1,500	0.811141	0.010083	1,502	0.239333	0.018095	500	0.114198	0.010813	864	0.347	0.010613	2,000
qwen2.5-7b	11	0.457949	7	0.702433	0.004566	10,000	0.300667	0.020506	500	0.698222	0.011839	1,500	0.777186	0.010736	1,502	0.297333	0.019018	500	0.143133	0.008419	864	0.286667	0.010095	2,000
granite3.2-8b	12	0.433736	7	0.674533	0.004681	10,000	0.288	0.020272	500	0.676	0.012067	1,500	0.810697	0.010097	1,502	0.126	0.012724	500	0.14892	0.01202	864	0.312	0.010325	2,000
gemma2-9b	13	0.433625	7	0.693833	0.004607	10,000	0.289333	0.020277	500	0.698222	0.011844	1,500	0.803817	0.01024	1,502	0.151333	0.014921	500	0.11767	0.010735	864	0.281167	0.01003	2,000
qwen2.5-coder-7b	14	0.421055	7	0.669233	0.004699	10,000	0.294	0.019954	500	0.670222	0.012114	1,500	0.650688	0.01227	1,502	0.206667	0.01688	500	0.178241	0.011168	864	0.278333	0.009999	2,000
mistral-nemo-12b	15	0.418227	7	0.642167	0.004781	10,000	0.278	0.019922	500	0.661111	0.012201	1,500	0.703506	0.011751	1,502	0.105333	0.01205	500	0.141975	0.010281	864	0.3955	0.010901	2,000
granite3.3-8b	16	0.414773	7	0.657767	0.00474	10,000	0.281333	0.020063	500	0.661778	0.012211	1,500	0.757878	0.011039	1,502	0.143333	0.013815	500	0.133488	0.01068	864	0.267833	0.009879	2,000
qwen2.5-3b	17	0.413322	7	0.663833	0.004683	10,000	0.272	0.018313	500	0.671333	0.012034	1,500	0.72814	0.011445	1,502	0.137333	0.013564	500	0.138117	0.008507	864	0.2825	0.010023	2,000
phi-3.5-mini	18	0.407997	7	0.6216	0.004838	10,000	0.288667	0.020219	500	0.633778	0.012408	1,500	0.755881	0.011034	1,502	0.162	0.015084	500	0.118056	0.010625	864	0.276	0.009959	2,000
internlm2.5-20b	19	0.405911	7	0.672067	0.004683	10,000	0.256	0.018935	500	0.623111	0.012465	1,500	0.701065	0.011762	1,502	0.126	0.013306	500	0.156636	0.010222	864	0.3065	0.010279	2,000
phi-4-mini	20	0.405322	7	0.648867	0.004751	10,000	0.259333	0.019368	500	0.699556	0.011791	1,500	0.721261	0.011531	1,502	0.223333	0.016543	500	0.026235	0.003987	864	0.258667	0.009729	2,000
internlm2.5-7b	21	0.403777	7	0.669	0.004684	10,000	0.301333	0.020366	500	0.654667	0.012237	1,500	0.738127	0.011274	1,502	0.088	0.010701	500	0.121142	0.008242	864	0.254167	0.009715	2,000
falcon3-3b	22	0.37716	7	0.608233	0.004876	10,000	0.212667	0.017999	500	0.550444	0.012833	1,500	0.630271	0.012436	1,502	0.23	0.017743	500	0.131173	0.008955	864	0.277333	0.009982	2,000
gemma2-2b	23	0.353013	7	0.6051	0.004889	10,000	0.272	0.01992	500	0.63	0.01247	1,500	0.736352	0.011373	1,502	0.016667	0.005653	500	0.003472	0.002002	864	0.2075	0.00907	2,000
qwen2.5-1.5b	24	0.351999	7	0.6102	0.004877	10,000	0.274	0.019966	500	0.619333	0.012541	1,500	0.567244	0.012788	1,502	0.041333	0.007072	500	0.111883	0.00862	864	0.24	0.009503	2,000
mixtral-8x7b	25	0.34896	7	0.487833	0.004979	10,000	0.198667	0.017071	500	0.566889	0.012736	1,500	0.652241	0.012229	1,502	0.1	0.010573	500	0.099923	0.007338	864	0.337167	0.010521	2,000
mistral-7b	26	0.310339	7	0.4532	0.004965	10,000	0.188	0.017363	500	0.468667	0.012835	1,500	0.592321	0.012614	1,502	0.060667	0.008712	500	0.122685	0.009192	864	0.286833	0.01011	2,000
gemma3-1b	27	0.272702	7	0.4567	0.004981	10,000	0.206	0.018105	500	0.526	0.012897	1,500	0.591212	0.012689	1,502	0.032	0.007879	500	0	0	864	0.097	0.006619	2,000
qwen2.5-0.5b	28	0.24309	7	0.464233	0.004927	10,000	0.18	0.015132	500	0.478889	0.012822	1,500	0.366844	0.012344	1,502	0.026	0.005582	500	0	0	864	0.185667	0.008607	2,000
falcon3-1b	29	0.22553	7	0.369567	0.004827	10,000	0.114	0.014227	500	0.44	0.012821	1,500	0.38482	0.012559	1,502	0.039333	0.007126	500	0.084491	0.008221	864	0.1465	0.007891	2,000
falcon3-7b	30	0.178545	7	0.1612	0.003662	10,000	0.063333	0.01078	500	0.176222	0.009786	1,500	0.088549	0.007245	1,502	0.316667	0.01995	500	0.158179	0.009973	864	0.285667	0.010098	2,000
command-r-35b	31	0.172135	3	null	null	null	null	null	null	null	null	null	null	null	null	0.055333	0.010059	500	0.123071	0.010913	864	0.338	0.010567	2,000

Open Telco Leaderboard Scores

Leaderboard scores extracted from Inspect evaluation logs for 31 models.

This dataset currently publishes scores only (no energy metrics).

Files

leaderboard_scores.csv: Flat table for the dataset viewer.
leaderboard_scores.json: Structured JSON with per-model benchmark scores, stderr, sample counts, and source eval file paths.

Schema (`leaderboard_scores.csv`)

Core columns:

model
rank
average
benchmarks_completed

Per-benchmark columns (for each benchmark):

<benchmark>_score
<benchmark>_stderr
<benchmark>_n

Benchmarks:

teleqna
teletables
oranbench
srsranbench
telemath
telelogs
three_gpp

Usage

from datasets import load_dataset

ds = load_dataset("GSMA/leaderboard", split="train")
print(ds.column_names)
print(ds[0])

Downloads last month: 188

Size of downloaded dataset files:

7.11 kB

Size of the auto-converted Parquet files:

16.1 kB

Number of rows:

Open Telco Leaderboard Scores

Files

Schema (leaderboard_scores.csv)

Usage

Schema (`leaderboard_scores.csv`)