Benchmark Almanac

TeleTables

TSLAM-150B 70% tops its leaderboard Gemini 3 Flash 48% leads the models we track
Not in model cards

Models answer questions about tables from 3GPP specifications, both from memory and with the table available.

The authors evaluated 20 models on 500 human-verified questions.

Made by
Anas Ezzakri and colleagues
Released
January 2026
Kind
Expert-written questions
Level
Expert
Scored
Checked automatically
Answers
Multiple choice
Tasks
All public
Measures
Capability
Language
English
In and out
Text, Tables → Text
Interaction
Single turn
Items from
Real work and use
Refresh
One fixed set
Run by
Anyone
Strongest model tested
Claude Opus 4.6 #39 on the capability index
Last model added
February 2026
Scores refreshed
November 20th, 2026
90
Entries on GSMA Open Telco Leaderboard’s leaderboard, 69 models we track

TSLAM-150B tops the leaderboard with 70% 2 models share the top score of 48% among models released by February 2026

Best score per entry from GSMA Open Telco Leaderboard’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 Gemini 3 Flash, at 48%. This view shows only models we track, each at its best.

The newest model with a score here came out in February 2026, and nobody has tested a newer one. Among the 20 highest-ranked models on the capability index, these have no score yet:

90 entries. Scroll sideways for the rest.

  1. 70.0 TSLAM-150B
  2. 61.8 OTel-LLM-8.3B-QnA
  3. 53.0 CHT-TeleX-8B
  4. 48.0 Gemini 3 Flash
  5. 48.0 Gemini 3.1 Pro
  6. 46.6 TeleLLM
  7. 45.0 Gemini 3 Pro
  8. 44.7 LTM
  9. 44.0 Claude Sonnet 4.6
  10. 43.0 Claude Opus 4.6
  11. 42.0 Kimi K2.5
  12. 41.0 GPT-5 nano
  13. 40.0 o4-mini
  14. 39.0 Claude Opus 4.5
  15. 38.0 Claude 3.7 Sonnet
  16. 38.0 Grok-3 mini
  17. 38.0 Gemini 2.5 Flash
  18. 37.0 GPT-4o mini
  19. 37.0 Gemini 2.0 Flash
  20. 37.0 Qwen3-235B-A22B-Instruct
  21. 37.0 GPT-5
  22. 36.2 netLLM
  23. 36.0 Claude Opus 4
  24. 36.0 Claude Opus 4.1
  25. 36.0 GPT-5.1
  26. 35.0 telco-slm-4B
  27. 35.0 GPT-4.1
  28. 35.0 Grok 4 Fast
  29. 35.0 MiniMax-M2.5
  30. 34.0 TSLAM-30B
  31. 34.0 Llama 4 Maverick
  32. 34.0 o3
  33. 34.0 Qwen3-32B
  34. 33.3 Gemma 3 27B
  35. 33.0 GPT-5.2
  36. 32.0 grok-code-fast-1
  37. 32.0 Claude Sonnet 4.5
  38. 32.0 Grok 4.1 Fast
  39. 31.0 Claude Sonnet 4
  40. 30.1 internlm2.5-7b
  41. 30.1 Qwen2.5-7B
  42. 30.0 Claude 3.5 Sonnet
  43. 30.0 GPT-4.1 mini
  44. 30.0 Gemini 2.5 Pro
  45. 30.0 gpt-oss-120b
  46. 29.9 Phi-4
  47. 29.7 mistral-small-24b
  48. 29.4 Qwen2.5-Coder (7B)
  49. 29.4 Qwen2.5-14B
  50. 29.2 Qwen2.5-72B
  51. 29.0 Llama 3.3 70B Instruct
  52. 29.0 o1
  53. 29.0 Cohere Command A
  54. 29.0 Qwen3-8B
  55. 29.0 MiMo-V2-Flash
  56. 28.9 Gemma 2 9B
  57. 28.9 Phi-3.5-mini
  58. 28.8 granite3.2-8b
  59. 28.5 Qwen2.5-32B
  60. 28.4 mistral-small-22b
  61. 28.1 granite3.3-8b
  62. 28.0 DeepSeek-V3
  63. 28.0 GPT-4.1 nano
  64. 28.0 GPT-5 mini
  65. 28.0 Claude Haiku 4.5
  66. 27.8 mistral-nemo-12b
  67. 27.7 falcon3-10b
  68. 27.4 Qwen2.5-1.5B
  69. 27.4 Gemma 3 12B
  70. 27.3 Gemma 3 4B
  71. 27.3 Gemma 2 27B
  72. 27.2 gemma2-2b
  73. 27.2 Qwen2.5-3B
  74. 27.0 Claude 3 Haiku
  75. 25.9 Phi-4 Mini
  76. 25.6 internlm2.5-20b
  77. 25.0 Claude 3.5 Haiku
  78. 21.3 falcon3-3b
  79. 21.0 GPT-3.5 Turbo
  80. 20.6 Gemma 3 1B
  81. 20.0 Command R7B
  82. 19.9 Mixtral 8x7B
  83. 18.8 Mistral 7B v0.1
  84. 18.0 GPT-4
  85. 18.0 Qwen2.5-0.5B
  86. 11.4 falcon3-1b
  87. 11.0 DeepSeek-R1
  88. 8.00 DeepSeek-V3.2
  89. 7.33 deepseek-r1-32b
  90. 6.33 falcon3-7b
  1. 1 TSLAM-150B 70%
  2. 2 OTel-LLM-8.3B-QnA 61.8%
  3. 3 CHT-TeleX-8B 53%
  4. 4 Gemini 3 Flash 48%
  5. 5 Gemini 3.1 Pro 48%
  6. 6 TeleLLM 46.6%
  7. 7 Gemini 3 Pro 45%
  8. 8 LTM 44.7%
  9. 9 Claude Sonnet 4.6 44%
  10. 10 Claude Opus 4.6 43%
  11. 11 Kimi K2.5 42%
  12. 12 GPT-5 nano 41%
  13. 13 o4-mini 40%
  14. 14 Claude Opus 4.5 39%
  15. 15 Claude 3.7 Sonnet 38%
  16. 16 Grok-3 mini 38%
  17. 17 Gemini 2.5 Flash Sep 2025 38%
  18. 18 GPT-4o mini 37%
  19. 19 Gemini 2.0 Flash Feb 2025 37%
  20. 20 Qwen3-235B-A22B-Instruct Jul 2025 37%
  21. 21 GPT-5 37%
  22. 22 netLLM 36.2%
  23. 23 Claude Opus 4 36%
  24. 24 Claude Opus 4.1 36%
  25. 25 GPT-5.1 36%
  26. 26 telco-slm-4B 35%
  27. 27 GPT-4.1 35%
  28. 28 Grok 4 Fast 35%
  29. 29 MiniMax-M2.5 35%
  30. 30 TSLAM-30B 34%
  31. 31 Llama 4 Maverick 34%
  32. 32 o3 34%
  33. 33 Qwen3-32B 34%
  34. 34 Gemma 3 27B 33.3%
  35. 35 GPT-5.2 33%
  36. 36 grok-code-fast-1 32%
  37. 37 Claude Sonnet 4.5 32%
  38. 38 Grok 4.1 Fast 32%
  39. 39 Claude Sonnet 4 31%
  40. 40 internlm2.5-7b 30.1%
  41. 41 Qwen2.5-7B 30.1%
  42. 42 Claude 3.5 Sonnet 30%
  43. 43 GPT-4.1 mini 30%
  44. 44 Gemini 2.5 Pro Jun 2025 30%
  45. 45 gpt-oss-120b 30%
  46. 46 Phi-4 29.9%
  47. 47 mistral-small-24b 29.7%
  48. 48 Qwen2.5-Coder (7B) 29.4%
  49. 49 Qwen2.5-14B 29.4%
  50. 50 Qwen2.5-72B 29.2%
  51. 51 Llama 3.3 70B Instruct 29%
  52. 52 o1 29%
  53. 53 Cohere Command A 29%
  54. 54 Qwen3-8B 29%
  55. 55 MiMo-V2-Flash 29%
  56. 56 Gemma 2 9B 28.9%
  57. 57 Phi-3.5-mini 28.9%
  58. 58 granite3.2-8b 28.8%
  59. 59 Qwen2.5-32B 28.5%
  60. 60 mistral-small-22b 28.4%
  61. 61 granite3.3-8b 28.1%
  62. 62 DeepSeek-V3 Mar 2025 28%
  63. 63 GPT-4.1 nano 28%
  64. 64 GPT-5 mini 28%
  65. 65 Claude Haiku 4.5 28%
  66. 66 mistral-nemo-12b 27.8%
  67. 67 falcon3-10b 27.7%
  68. 68 Qwen2.5-1.5B 27.4%
  69. 69 Gemma 3 12B 27.4%
  70. 70 Gemma 3 4B 27.3%
  71. 71 Gemma 2 27B 27.3%
  72. 72 gemma2-2b 27.2%
  73. 73 Qwen2.5-3B 27.2%
  74. 74 Claude 3 Haiku 27%
  75. 75 Phi-4 Mini 25.9%
  76. 76 internlm2.5-20b 25.6%
  77. 77 Claude 3.5 Haiku 25%
  78. 78 falcon3-3b 21.3%
  79. 79 GPT-3.5 Turbo Jan 2024 21%
  80. 80 Gemma 3 1B 20.6%
  81. 81 Command R7B 20%
  82. 82 Mixtral 8x7B 19.9%
  83. 83 Mistral 7B v0.1 18.8%
  84. 84 GPT-4 Jun 2023 18%
  85. 85 Qwen2.5-0.5B 18%
  86. 86 falcon3-1b 11.4%
  87. 87 DeepSeek-R1 May 2025 11%
  88. 88 DeepSeek-V3.2 8.00%
  89. 89 deepseek-r1-32b 7.33%
  90. 90 falcon3-7b 6.33%

Source: GSMA Open Telco Leaderboard’s leaderboard

  1. 1 Gemini 3 Flash 48%
  2. 2 Gemini 3.1 Pro 48%
  3. 3 Gemini 3 Pro 45%
  4. 4 Claude Sonnet 4.6 44%
  5. 5 Claude Opus 4.6 43%
  6. 6 Kimi K2.5 42%
  7. 7 GPT-5 nano 41%
  8. 8 o4-mini 40%
  9. 9 Claude Opus 4.5 39%
  10. 10 Claude 3.7 Sonnet 38%
  11. 11 Grok-3 mini 38%
  12. 12 Gemini 2.5 Flash Sep 2025 38%
  13. 13 GPT-4o mini 37%
  14. 14 Gemini 2.0 Flash Feb 2025 37%
  15. 15 Qwen3-235B-A22B-Instruct Jul 2025 37%
  16. 16 GPT-5 37%
  17. 17 Claude Opus 4 36%
  18. 18 Claude Opus 4.1 36%
  19. 19 GPT-5.1 36%
  20. 20 GPT-4.1 35%
  21. 21 Grok 4 Fast 35%
  22. 22 MiniMax-M2.5 35%
  23. 23 Llama 4 Maverick 34%
  24. 24 o3 34%
  25. 25 Qwen3-32B 34%
  26. 26 Gemma 3 27B 33.3%
  27. 27 GPT-5.2 33%
  28. 28 grok-code-fast-1 32%
  29. 29 Claude Sonnet 4.5 32%
  30. 30 Grok 4.1 Fast 32%
  31. 31 Claude Sonnet 4 31%
  32. 32 Qwen2.5-7B 30.1%
  33. 33 Claude 3.5 Sonnet 30%
  34. 34 GPT-4.1 mini 30%
  35. 35 Gemini 2.5 Pro Jun 2025 30%
  36. 36 gpt-oss-120b 30%
  37. 37 Phi-4 29.9%
  38. 38 Qwen2.5-Coder (7B) 29.4%
  39. 39 Qwen2.5-14B 29.4%
  40. 40 Qwen2.5-72B 29.2%
  41. 41 Llama 3.3 70B Instruct 29%
  42. 42 o1 29%
  43. 43 Cohere Command A 29%
  44. 44 Qwen3-8B 29%
  45. 45 MiMo-V2-Flash 29%
  46. 46 Gemma 2 9B 28.9%
  47. 47 Phi-3.5-mini 28.9%
  48. 48 Qwen2.5-32B 28.5%
  49. 49 DeepSeek-V3 Mar 2025 28%
  50. 50 GPT-4.1 nano 28%
  51. 51 GPT-5 mini 28%
  52. 52 Claude Haiku 4.5 28%
  53. 53 Qwen2.5-1.5B 27.4%
  54. 54 Gemma 3 12B 27.4%
  55. 55 Gemma 3 4B 27.3%
  56. 56 Gemma 2 27B 27.3%
  57. 57 Qwen2.5-3B 27.2%
  58. 58 Claude 3 Haiku 27%
  59. 59 Phi-4 Mini 25.9%
  60. 60 Claude 3.5 Haiku 25%
  61. 61 GPT-3.5 Turbo Jan 2024 21%
  62. 62 Gemma 3 1B 20.6%
  63. 63 Command R7B 20%
  64. 64 Mixtral 8x7B 19.9%
  65. 65 Mistral 7B v0.1 18.8%
  66. 66 GPT-4 Jun 2023 18%
  67. 67 Qwen2.5-0.5B 18%
  68. 68 DeepSeek-R1 May 2025 11%
  69. 69 DeepSeek-V3.2 8.00%

Source: GSMA Open Telco Leaderboard’s leaderboard

#EntryScoreSettingDateSource
1 TSLAM-150B 70% Official leaderboard
2 OTel-LLM-8.3B-QnA 61.8% Official leaderboard
3 CHT-TeleX-8B 53% Official leaderboard
4 Gemini 3 Flash 48% Dec 2025 Official leaderboard
5 Gemini 3.1 Pro 48% Feb 2026 Official leaderboard
6 TeleLLM 46.6% Official leaderboard
7 Gemini 3 Pro 45% Nov 2025 Official leaderboard
8 LTM 44.7% Official leaderboard
9 Claude Sonnet 4.6 44% Feb 2026 Official leaderboard
10 Claude Opus 4.6 43% Feb 2026 Official leaderboard
11 Kimi K2.5 42% Jan 2026 Official leaderboard
12 GPT-5 nano 41% Aug 2025 Official leaderboard
13 o4-mini 40% Apr 2025 Official leaderboard
14 Claude Opus 4.5 39% Nov 2025 Official leaderboard
15 Claude 3.7 Sonnet 38% Feb 2025 Official leaderboard
16 Grok-3 mini 38% Apr 2025 Official leaderboard
17 Gemini 2.5 Flash Sep 2025 38% Sep 2025 Official leaderboard
18 GPT-4o mini 37% Jul 2024 Official leaderboard
19 Gemini 2.0 Flash Feb 2025 37% Feb 2025 Official leaderboard
20 Qwen3-235B-A22B-Instruct Jul 2025 37% Jul 2025 Official leaderboard
21 GPT-5 37% Aug 2025 Official leaderboard
22 netLLM 36.2% Official leaderboard
23 Claude Opus 4 36% May 2025 Official leaderboard
24 Claude Opus 4.1 36% Aug 2025 Official leaderboard
25 GPT-5.1 36% Nov 2025 Official leaderboard
26 telco-slm-4B 35% Official leaderboard
27 GPT-4.1 35% Apr 2025 Official leaderboard
28 Grok 4 Fast 35% Sep 2025 Official leaderboard
29 MiniMax-M2.5 35% Feb 2026 Official leaderboard
30 TSLAM-30B 34% Official leaderboard
31 Llama 4 Maverick 34% Apr 2025 Official leaderboard
32 o3 34% Apr 2025 Official leaderboard
33 Qwen3-32B 34% Apr 2025 Official leaderboard
34 Gemma 3 27B 33.3% Mar 2025 Official leaderboard
35 GPT-5.2 33% Dec 2025 Official leaderboard
36 grok-code-fast-1 32% Aug 2025 Official leaderboard
37 Claude Sonnet 4.5 32% Sep 2025 Official leaderboard
38 Grok 4.1 Fast 32% Nov 2025 Official leaderboard
39 Claude Sonnet 4 31% May 2025 Official leaderboard
40 internlm2.5-7b 30.1% Official leaderboard
41 Qwen2.5-7B 30.1% Sep 2024 Official leaderboard
42 Claude 3.5 Sonnet 30% Jun 2024 Official leaderboard
43 GPT-4.1 mini 30% Apr 2025 Official leaderboard
44 Gemini 2.5 Pro Jun 2025 30% Jun 2025 Official leaderboard
45 gpt-oss-120b 30% Aug 2025 Official leaderboard
46 Phi-4 29.9% Dec 2024 Official leaderboard
47 mistral-small-24b 29.7% Official leaderboard
48 Qwen2.5-Coder (7B) 29.4% Sep 2024 Official leaderboard
49 Qwen2.5-14B 29.4% Sep 2024 Official leaderboard
50 Qwen2.5-72B 29.2% Sep 2024 Official leaderboard
51 Llama 3.3 70B Instruct 29% Dec 2024 Official leaderboard
52 o1 29% Dec 2024 Official leaderboard
53 Cohere Command A 29% Mar 2025 Official leaderboard
54 Qwen3-8B 29% Apr 2025 Official leaderboard
55 MiMo-V2-Flash 29% Dec 2025 Official leaderboard
56 Gemma 2 9B 28.9% Jun 2024 Official leaderboard
57 Phi-3.5-mini 28.9% Aug 2024 Official leaderboard
58 granite3.2-8b 28.8% Official leaderboard
59 Qwen2.5-32B 28.5% Sep 2024 Official leaderboard
60 mistral-small-22b 28.4% Official leaderboard
61 granite3.3-8b 28.1% Official leaderboard
62 DeepSeek-V3 Mar 2025 28% Mar 2025 Official leaderboard
63 GPT-4.1 nano 28% Apr 2025 Official leaderboard
64 GPT-5 mini 28% Aug 2025 Official leaderboard
65 Claude Haiku 4.5 28% Oct 2025 Official leaderboard
66 mistral-nemo-12b 27.8% Official leaderboard
67 falcon3-10b 27.7% Official leaderboard
68 Qwen2.5-1.5B 27.4% Sep 2024 Official leaderboard
69 Gemma 3 12B 27.4% Mar 2025 Official leaderboard
70 Gemma 3 4B 27.3% Mar 2025 Official leaderboard
71 Gemma 2 27B 27.3% Jun 2024 Official leaderboard
72 gemma2-2b 27.2% Official leaderboard
73 Qwen2.5-3B 27.2% Sep 2024 Official leaderboard
74 Claude 3 Haiku 27% Mar 2024 Official leaderboard
75 Phi-4 Mini 25.9% Feb 2025 Official leaderboard
76 internlm2.5-20b 25.6% Official leaderboard
77 Claude 3.5 Haiku 25% Oct 2024 Official leaderboard
78 falcon3-3b 21.3% Official leaderboard
79 GPT-3.5 Turbo Jan 2024 21% Jan 2024 Official leaderboard
80 Gemma 3 1B 20.6% Mar 2025 Official leaderboard
81 Command R7B 20% Dec 2024 Official leaderboard
82 Mixtral 8x7B 19.9% Dec 2023 Official leaderboard
83 Mistral 7B v0.1 18.8% Sep 2023 Official leaderboard
84 GPT-4 Jun 2023 18% Jun 2023 Official leaderboard
85 Qwen2.5-0.5B 18% Sep 2024 Official leaderboard
86 falcon3-1b 11.4% Official leaderboard
87 DeepSeek-R1 May 2025 11% May 2025 Official leaderboard
88 DeepSeek-V3.2 8.00% Dec 2025 Official leaderboard
89 deepseek-r1-32b 7.33% Official leaderboard
90 falcon3-7b 6.33% Official leaderboard
#ModelScoreSettingDateSource
1 Gemini 3 Flash 48% Dec 2025 Official leaderboard
2 Gemini 3.1 Pro 48% Feb 2026 Official leaderboard
3 Gemini 3 Pro 45% Nov 2025 Official leaderboard
4 Claude Sonnet 4.6 44% Feb 2026 Official leaderboard
5 Claude Opus 4.6 43% Feb 2026 Official leaderboard
6 Kimi K2.5 42% Jan 2026 Official leaderboard
7 GPT-5 nano 41% Aug 2025 Official leaderboard
8 o4-mini 40% Apr 2025 Official leaderboard
9 Claude Opus 4.5 39% Nov 2025 Official leaderboard
10 Claude 3.7 Sonnet 38% Feb 2025 Official leaderboard
11 Grok-3 mini 38% Apr 2025 Official leaderboard
12 Gemini 2.5 Flash Sep 2025 38% Sep 2025 Official leaderboard
13 GPT-4o mini 37% Jul 2024 Official leaderboard
14 Gemini 2.0 Flash Feb 2025 37% Feb 2025 Official leaderboard
15 Qwen3-235B-A22B-Instruct Jul 2025 37% Jul 2025 Official leaderboard
16 GPT-5 37% Aug 2025 Official leaderboard
17 Claude Opus 4 36% May 2025 Official leaderboard
18 Claude Opus 4.1 36% Aug 2025 Official leaderboard
19 GPT-5.1 36% Nov 2025 Official leaderboard
20 GPT-4.1 35% Apr 2025 Official leaderboard
21 Grok 4 Fast 35% Sep 2025 Official leaderboard
22 MiniMax-M2.5 35% Feb 2026 Official leaderboard
23 Llama 4 Maverick 34% Apr 2025 Official leaderboard
24 o3 34% Apr 2025 Official leaderboard
25 Qwen3-32B 34% Apr 2025 Official leaderboard
26 Gemma 3 27B 33.3% Mar 2025 Official leaderboard
27 GPT-5.2 33% Dec 2025 Official leaderboard
28 grok-code-fast-1 32% Aug 2025 Official leaderboard
29 Claude Sonnet 4.5 32% Sep 2025 Official leaderboard
30 Grok 4.1 Fast 32% Nov 2025 Official leaderboard
31 Claude Sonnet 4 31% May 2025 Official leaderboard
32 Qwen2.5-7B 30.1% Sep 2024 Official leaderboard
33 Claude 3.5 Sonnet 30% Jun 2024 Official leaderboard
34 GPT-4.1 mini 30% Apr 2025 Official leaderboard
35 Gemini 2.5 Pro Jun 2025 30% Jun 2025 Official leaderboard
36 gpt-oss-120b 30% Aug 2025 Official leaderboard
37 Phi-4 29.9% Dec 2024 Official leaderboard
38 Qwen2.5-Coder (7B) 29.4% Sep 2024 Official leaderboard
39 Qwen2.5-14B 29.4% Sep 2024 Official leaderboard
40 Qwen2.5-72B 29.2% Sep 2024 Official leaderboard
41 Llama 3.3 70B Instruct 29% Dec 2024 Official leaderboard
42 o1 29% Dec 2024 Official leaderboard
43 Cohere Command A 29% Mar 2025 Official leaderboard
44 Qwen3-8B 29% Apr 2025 Official leaderboard
45 MiMo-V2-Flash 29% Dec 2025 Official leaderboard
46 Gemma 2 9B 28.9% Jun 2024 Official leaderboard
47 Phi-3.5-mini 28.9% Aug 2024 Official leaderboard
48 Qwen2.5-32B 28.5% Sep 2024 Official leaderboard
49 DeepSeek-V3 Mar 2025 28% Mar 2025 Official leaderboard
50 GPT-4.1 nano 28% Apr 2025 Official leaderboard
51 GPT-5 mini 28% Aug 2025 Official leaderboard
52 Claude Haiku 4.5 28% Oct 2025 Official leaderboard
53 Qwen2.5-1.5B 27.4% Sep 2024 Official leaderboard
54 Gemma 3 12B 27.4% Mar 2025 Official leaderboard
55 Gemma 3 4B 27.3% Mar 2025 Official leaderboard
56 Gemma 2 27B 27.3% Jun 2024 Official leaderboard
57 Qwen2.5-3B 27.2% Sep 2024 Official leaderboard
58 Claude 3 Haiku 27% Mar 2024 Official leaderboard
59 Phi-4 Mini 25.9% Feb 2025 Official leaderboard
60 Claude 3.5 Haiku 25% Oct 2024 Official leaderboard
61 GPT-3.5 Turbo Jan 2024 21% Jan 2024 Official leaderboard
62 Gemma 3 1B 20.6% Mar 2025 Official leaderboard
63 Command R7B 20% Dec 2024 Official leaderboard
64 Mixtral 8x7B 19.9% Dec 2023 Official leaderboard
65 Mistral 7B v0.1 18.8% Sep 2023 Official leaderboard
66 GPT-4 Jun 2023 18% Jun 2023 Official leaderboard
67 Qwen2.5-0.5B 18% Sep 2024 Official leaderboard
68 DeepSeek-R1 May 2025 11% May 2025 Official leaderboard
69 DeepSeek-V3.2 8.00% Dec 2025 Official leaderboard

Number one over time

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

Google DeepMind
GSMA Open Telco Leaderboard leaderboard Not tested since Feb 2026

Source: GSMA Open Telco Leaderboard’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 GSMA Open Telco Leaderboard’s leaderboard Best so far
0% 20% 40% 60% 80% 100% 2023 2024 2025 2026 Near the ceiling (85%) Gemini 3 Flash GPT-5 nano GPT-4o mini Claude 3.5 Sonnet GPT-3.5 Turbo GPT-4 0% 20% 40% 60% 80% 100% 2023 2024 2025 2026 Near the ceiling (85%) Gemini 3 Flash GPT-5 nano

Source: GSMA Open Telco Leaderboard’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 Knowledge

All 332 in Knowledge
  • MMLU-Pro In use Harder 10-option version of MMLU with reasoning-focused questions across 14 subjects. Intern-S2-397B 89.8%
  • MMLU In use 57-subject multiple-choice exam questions from elementary to professional level. Scored by accuracy. Seed1.8 92.3%
  • SuperGPQA In use 26,529 graduate-level multiple-choice questions across 285 disciplines. Accuracy. Qwen 3.6 Max (Preview) 73.9%
  • WideSearch In use 200 bilingual broad info-seeking tasks: agent collects many atomic facts into a table. Scored by success rate and row/item F1. Tencent Hy4 preview 83.9%
  • BBEH micro average In use Models solve harder replacements for each of the 23 BIG-Bench Hard tasks, 4,520 examples in total. The score is the share of all examples answered correctly. Gemma 4 31B IT 74.4%
  • HLE-Diamond In use Models answer 1,000 cleaned expert questions from Humanity's Last Exam, 500 on reasoning and 500 on knowledge, closed book. Graded on exact answers. Claude Fable 5.2 77.9%

Data audit

9 of 10 checks pass, 1 warning. Look first at the Newest score check. The newest came 9 months ago, in February 2026, for Gemini 3.1 Pro. Every benchmark

Catalog record 2 pass
  • Pass Task access All tasks are public.
  • Pass Confidence Checked against a primary source.
Scores 1 warning 3 pass 1 note
  • Pass Independent runs Scores from the official leaderboard (90 entries).
  • Pass Models scored 69 models we track, plus 21 other entries on its leaderboard.
  • Warning Newest score The newest came 9 months ago, in February 2026, for Gemini 3.1 Pro.
  • Note Frontier models None of the top 10 has a score. The strongest tested is Claude Opus 4.6 (#39).
  • Pass Headroom The best is 48%, under 85%, so it still separates models.
Leaderboard import 3 pass
  • Pass Importer We read the board on GSMA Open Telco Leaderboard: 70 rows for models we track and 21 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