Benchmark Almanac

CNN/DailyMail Summarization

Cohere Command beta (52.4B) 0.16 tops its leaderboard OPT-175B 0.15 leads the models we track
Not in model cards

Models summarize news articles from CNN and the Daily Mail. The harness compares generated summaries with the publishers’ highlights.

Released
June 2015
Kind
Faithful to sources
Level
School
Scored
Checked automatically
Answers
Open-ended
Tasks
All public
Measures
Capability
Language
English
In and out
Text → Text
Interaction
Single turn
Items from
Real work and use
Refresh
One fixed set
Run by
Anyone
Last model added
July 2022
Scores refreshed
November 20th, 2026
42
Entries on HELM Classic’s leaderboard, 6 models we track

Cohere Command beta (52.4B) tops the leaderboard with 0.16 OPT-175B leads with 0.15 among models released by July 2022

Best score per entry from HELM Classic’s leaderboard, read on November 20th, 2026. Where a model ran with several settings or harnesses, we keep its best. Names in gray aren't in our model list: fine-tunes, agent systems, products and older models. They don't count toward leaders. The best model we track is OPT-175B, at 0.15. This view shows only models we track, each at its best.

The newest model with a score here came out in July 2022, and nobody has tested a newer one. Among the 20 highest-ranked models on the capability index, these have no score yet:

  1. 1 Cohere Command beta (52.4B) 0.16
  2. 2 TNLG v2 (530B) 0.16
  3. 3 text-davinci-003 0.16
  4. 4 Anthropic-LM v4-s3 (52B) 0.15
  5. 5 GLM (130B) 0.15
  6. 6 text-davinci-002 0.15
  7. 7 Cohere xlarge v20221108 (52.4B) 0.15
  8. 8 Cohere Command beta (6.1B) 0.15
  9. 9 InstructPalmyra (30B) 0.15
  10. 10 text-curie-001 0.15
  11. 11 text-babbage-001 0.15
  12. 12 Luminous Supreme (70B) 0.15
  13. 13 Jurassic-2 Jumbo (178B) 0.15
  14. 14 TNLG v2 (6.7B) 0.15
  15. 15 OPT-175B 0.15
  16. 16 J1-Grande v2 beta (17B) 0.15
  17. 17 Jurassic-2 Grande (17B) 0.14
  18. 18 J1-Jumbo v1 (178B) 0.14
  19. 19 Cohere xlarge v20220609 (52.4B) 0.14
  20. 20 J1-Grande v1 (17B) 0.14
  21. 21 Luminous Extended (30B) 0.14
  22. 22 OPT-66B 0.14
  23. 23 Jurassic-2 Large (7.5B) 0.14
  24. 24 text-ada-001 0.14
  25. 25 J1-Large v1 (7.5B) 0.13
  26. 26 GPT-J-6B 0.13
  27. 27 davinci (175B) 0.13
  28. 28 Cohere large v20220720 (13.1B) 0.13
  29. 29 GPT-NeoX-20B 0.12
  30. 30 T0pp (11B)☠ 0.12
  31. 31 Cohere medium v20221108 (6.1B) 0.12
  32. 32 curie (6.7B) 0.11
  33. 33 Luminous Base (13B) 0.11
  34. 34 ada (350M) 0.09
  35. 35 BLOOM-176B 0.08
  36. 36 babbage (1.3B) 0.08
  37. 37 Cohere medium v20220720 (6.1B) 0.08
  38. 38 Cohere small v20220720 (410M) 0.06
  39. 39 Palmyra X (43B) 0.05
  40. 40 T5-11B 0.04
  41. 41 UL2 (20B) 0.03
  42. 42 YaLM (100B) 0.02

Source: HELM Classic’s leaderboard

Source: HELM Classic’s leaderboard

#EntryScoreSettingDateSource
1 Cohere Command beta (52.4B) 0.16 Official leaderboard
2 TNLG v2 (530B) 0.16 Official leaderboard
3 text-davinci-003 0.16 Official leaderboard
4 Anthropic-LM v4-s3 (52B) 0.15 Official leaderboard
5 GLM (130B) 0.15 Official leaderboard
6 text-davinci-002 0.15 Official leaderboard
7 Cohere xlarge v20221108 (52.4B) 0.15 Official leaderboard
8 Cohere Command beta (6.1B) 0.15 Official leaderboard
9 InstructPalmyra (30B) 0.15 Official leaderboard
10 text-curie-001 0.15 Official leaderboard
11 text-babbage-001 0.15 Official leaderboard
12 Luminous Supreme (70B) 0.15 Official leaderboard
13 Jurassic-2 Jumbo (178B) 0.15 Official leaderboard
14 TNLG v2 (6.7B) 0.15 Official leaderboard
15 OPT-175B 0.15 May 2022 Official leaderboard
16 J1-Grande v2 beta (17B) 0.15 Official leaderboard
17 Jurassic-2 Grande (17B) 0.14 Official leaderboard
18 J1-Jumbo v1 (178B) 0.14 Official leaderboard
19 Cohere xlarge v20220609 (52.4B) 0.14 Official leaderboard
20 J1-Grande v1 (17B) 0.14 Official leaderboard
21 Luminous Extended (30B) 0.14 Official leaderboard
22 OPT-66B 0.14 May 2022 Official leaderboard
23 Jurassic-2 Large (7.5B) 0.14 Official leaderboard
24 text-ada-001 0.14 Official leaderboard
25 J1-Large v1 (7.5B) 0.13 Official leaderboard
26 GPT-J-6B 0.13 Aug 2021 Official leaderboard
27 davinci (175B) 0.13 Official leaderboard
28 Cohere large v20220720 (13.1B) 0.13 Official leaderboard
29 GPT-NeoX-20B 0.12 Apr 2022 Official leaderboard
30 T0pp (11B)☠ 0.12 Official leaderboard
31 Cohere medium v20221108 (6.1B) 0.12 Official leaderboard
32 curie (6.7B) 0.11 Official leaderboard
33 Luminous Base (13B) 0.11 Official leaderboard
34 ada (350M) 0.09 Official leaderboard
35 BLOOM-176B 0.08 Jul 2022 Official leaderboard
36 babbage (1.3B) 0.08 Official leaderboard
37 Cohere medium v20220720 (6.1B) 0.08 Official leaderboard
38 Cohere small v20220720 (410M) 0.06 Official leaderboard
39 Palmyra X (43B) 0.05 Official leaderboard
40 T5-11B 0.04 Oct 2019 Official leaderboard
41 UL2 (20B) 0.03 Official leaderboard
42 YaLM (100B) 0.02 Official leaderboard
#ModelScoreSettingDateSource
1 OPT-175B 0.15 May 2022 Official leaderboard
2 OPT-66B 0.14 May 2022 Official leaderboard
3 GPT-J-6B 0.13 Aug 2021 Official leaderboard
4 GPT-NeoX-20B 0.12 Apr 2022 Official leaderboard
5 BLOOM-176B 0.08 Jul 2022 Official leaderboard
6 T5-11B 0.04 Oct 2019 Official leaderboard

Rankings by part

We read 3 parts from HELM Classic’s leaderboard: BERTScore (F1), SummaC and QAFactEval. Each ranks models on one part of the test. OPT-175B leads 1 of them, and 2 other models lead the rest. Arrows show how far a model moved from its overall rank among the models we track.

BERTScore (F1)

6 models we track have a score here.

  1. 1 OPT-175B 0.28
  2. 2 OPT-66B 0.26
  3. 3 GPT-J-6B 0.25
  4. 4 GPT-NeoX-20B 0.23
  5. 5 BLOOM-176B 0.08
  6. 6 T5-11B -0.17

Number one over time

Lab claims count from the day the card came out, independent runs from the model’s release.

Meta
HELM Classic leaderboard Not tested since Jul 2022
OPT-175B 0.15

Source: HELM Classic’s leaderboard

Scores over time

All the scores we have, one dot each, so a model tested at several settings appears more than once.

Plot with other benchmarks
On HELM Classic’s leaderboard
0.05 0.07 0.10 0.13 0.15 2020 2021 2022 0.05 0.07 0.10 0.13 0.15 2020 2021 2022

Source: HELM Classic’s leaderboard

No lab model card reports it yet

We read the headline results of 540 launch posts from 57 labs, and none of them list this benchmark.

More in Factuality

All 164 in Factuality
  • SimpleVQA In use Models answer 2,025 short factual questions about images across 9 tasks and 9 topics. An LLM judge grades answers against fixed references. Qwen3.7-Plus 81.7%
  • HallusionBench In use Models answer 1,129 yes or no questions about 346 images, built in pairs to expose language hallucination and visual illusion. Scored by accuracy. Keye-VL-2.0-30B-A3B 73.5%
  • SimpleQA Verified In use 1,000 human-verified, de-duplicated short-answer factual questions measuring parametric knowledge. Gemini 3.1 Pro 77.5%
  • AA-Omniscience v1.1 In use Models answer 3,000 knowledge questions in six domains without tools and may abstain. The index adds correct answers and subtracts wrong ones. Inkling 2 Small 16.1
  • GaRAGe In use Models write long-form RAG answers to 2,366 questions with human-annotated grounding passages. Scoring checks whether answers use the relevant passages. Amazon Nova Pro 60.7%
  • Kimi Hallu Bench In use Models answer hard prompts drawn from practical work tasks. Fact-checking agents score the share of claims that are hallucinated. GPT-6 Astra 4.20%

Data audit

8 of 9 checks pass, 1 warning. Look first at the Newest score check. The newest came 4 years ago, in July 2022, for BLOOM-176B. Every benchmark

Catalog record 2 pass
  • Pass Task access All tasks are public.
  • Pass Confidence Checked against a primary source.
Scores 1 warning 2 pass 2 notes
  • Pass Independent runs Scores from the official leaderboard (42 entries).
  • Pass Models scored 6 models we track, plus 36 other entries on its leaderboard.
  • Warning Newest score The newest came 4 years ago, in July 2022, for BLOOM-176B.
  • Note Frontier models None of its models are on the capabilities index.
  • Note Headroom Scored as a plain number, with no fixed ceiling.
Leaderboard import 3 pass
  • Pass Importer We read the board on HELM Classic: 6 rows for models we track and 36 other entries at the last read.
  • Pass Last read November 20th, 2026.
  • Pass Read worked The last read worked.
Consistency 1 pass 1 note
  • Note Lab claims No model has both a lab number and an independent run.
  • Pass Known issues No published audit found problems.

How we check