Benchmark Almanac

Pre-Flight

GPT-5.6 Sol 84% leads since July 2026
Not in model cards

Models answer questions about aviation regulations, dispatch, pilot procedures and airport operations. The evaluator compares each choice with the answer key.

The date uses the creation month of the published dataset repository.

Made by
Airside Labs
Released
March 2025
Kind
Other professions
Level
Expert
Scored
Checked automatically
Answers
Multiple choice
Tasks
All public
Measures
Capability
Language
English
In and out
Text → Text
Interaction
Single turn
Refresh
One fixed set
Run by
Anyone
Occupations
Transportation and material moving
In views
Law and finance
Strongest model tested
GPT-6 Astra #4 on the capability index
Last model added
September 2026
Scores refreshed
November 20th, 2026
75
Entries on its own leaderboard, 74 models we track

GPT-5.6 Sol leads with 84% among models released by September 2026 GPT-5.6 Sol leads with 84% among models released by September 2026

Best score per entry from the benchmark's own leaderboard, read on November 20th, 2026. Where a model ran with several settings or harnesses, we keep its best. Every run is in the next chart. Names in gray aren't in our model list: fine-tunes, agent systems, products and older models. They don't count toward leaders. This view shows only models we track, each at its best.

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

75 entries. Scroll sideways for the rest.

  1. 1 GPT-5.6 Sol 84%
  2. 2 GPT-5.6 Terra 83.7%
  3. 3 Kimi K3 83.6%
  4. 4 GPT-6 Astra 83.6%
  5. 5 Claude Opus 5.5 82.9%
  6. 6 GPT-5.5 82.7%
  7. 7 Gemini 3.8 Flash 82.5%
  8. 8 Gemini 3.7 Flash 82.2%
  9. 9 Claude Fable 5 82%
  10. 10 Claude Fable 5.1 81.9%
  11. 11 GPT-5.6 Luna 81.3%
  12. 12 Qwen 3.8 Max 81.2%
  13. 13 Claude Opus 5 80.5%
  14. 14 Gemini 4 Argon 80.5%
  15. 15 GPT-5 80.3%
  16. 16 Claude Opus 4.6 80.3%
  17. 17 Claude Opus 4.7 79.7%
  18. 18 DeepSeek V4 Pro 0813 79.5%
  19. 19 Grok 4.5 79.2%
  20. 20 Gemini 2.5 Pro Jun 2025 79%
  21. 21 Claude Opus 4.8 79%
  22. 22 Claude Sonnet 5.5 78.5%
  23. 23 Claude Sonnet 5 78.3%
  24. 24 Claude Opus 4.1 78%
  25. 25 GLM-5.3 77.8%
  26. 26 Qwen3.5-122B-A10B int4, local 2x Spark 77.3%
  27. 27 GPT-5.1 77%
  28. 28 Gemini 3.6 Flash 76.8%
  29. 29 Llama-3.1-Nemotron-70B-Instruct 76.8%
  30. 30 Gemma 4 31B IT 76.5%
  31. 31 Mistral Large 3 76.2%
  32. 32 Grok 4.6 76.2%
  33. 33 MiniMax-M3 75.8%
  34. 34 GPT-5.4 75.7%
  35. 35 Qwen 3.6 35B-A3B FP8, local Spark 75.7%
  36. 36 Gemini 3.5 Flash 75.7%
  37. 37 Claude Opus 4.5 75.3%
  38. 38 DeepSeek V4 Flash 0731 75.2%
  39. 39 Gemini 2.5 Flash Sep 2025 75%
  40. 40 GPT-5.4 Mini 75%
  41. 41 Llama 3.3 70B 74.7%
  42. 42 Claude 3.7 Sonnet Prior evaluation 74.7%
  43. 43 Claude Sonnet 4.5 74.7%
  44. 44 Llama 4 Scout 74.3%
  45. 45 GPT-5.2 74.3%
  46. 46 Qwen3-Coder-Next FP8, local Spark 73.7%
  47. 47 Solar Pro 4 73.5%
  48. 48 GPT-4o mini Prior evaluation 73.3%
  49. 49 GPT-4o Nov 2024, Prior evaluation 73.3%
  50. 50 Gemma 4 26B A4B local Spark 73.3%
  51. 51 Claude Sonnet 4.6 72.7%
  52. 52 Claude Haiku 4.5 72.3%
  53. 53 MiniMax-M2.7 AWQ, local 2x Spark 72.3%
  54. 54 Qwen3.5-35B-A3B FP8, local Spark 72%
  55. 55 Claude 3.5 Sonnet (October 2024) Prior evaluation 71.3%
  56. 56 gpt-oss-120b MXFP4, local Spark 71.3%
  57. 57 Nemotron 3 Super NVFP4, local Spark 71.3%
  58. 58 Qwen3.6 27B 71.2%
  59. 59 Llama 3-70B Groq, Prior evaluation 70.7%
  60. 60 Qwen3-32B 70.7%
  61. 61 Gemini 2.5 Flash-Lite Sep 2025 70.1%
  62. 62 DiffusionGemma 26B A4B NVFP4, local Spark 69.3%
  63. 63 Claude 3 Haiku Prior evaluation 68.3%
  64. 64 Nemotron 3 Nano 30B-A3B local Spark 68%
  65. 65 Qwen 3.8 27B 67.1%
  66. 66 Claude 3.5 Haiku Prior evaluation 66.7%
  67. 67 GPT-4 Turbo (Jan 2024) Prior evaluation 66%
  68. 68 Llama 3-8B Groq, Prior evaluation 66%
  69. 69 gpt-oss-safeguard-20b 65.3%
  70. 70 GPT-3.5 Turbo Jan 2024, Prior evaluation 64%
  71. 71 gpt-oss-20b 63.7%
  72. 72 Gemma 2 9B Prior evaluation 62.3%
  73. 73 QwQ-32B Prior evaluation 58.7%
  74. 74 Llama 3.1-8B 58%
  75. 75 ALLaM 2 7B 25.3%

Source: Official Pre-Flight leaderboard

74 models. Scroll sideways for the rest.

  1. 1 GPT-5.6 Sol 84%
  2. 2 GPT-5.6 Terra 83.7%
  3. 3 Kimi K3 83.6%
  4. 4 GPT-6 Astra 83.6%
  5. 5 Claude Opus 5.5 82.9%
  6. 6 GPT-5.5 82.7%
  7. 7 Gemini 3.8 Flash 82.5%
  8. 8 Gemini 3.7 Flash 82.2%
  9. 9 Claude Fable 5 82%
  10. 10 Claude Fable 5.1 81.9%
  11. 11 GPT-5.6 Luna 81.3%
  12. 12 Qwen 3.8 Max 81.2%
  13. 13 Claude Opus 5 80.5%
  14. 14 Gemini 4 Argon 80.5%
  15. 15 GPT-5 80.3%
  16. 16 Claude Opus 4.6 80.3%
  17. 17 Claude Opus 4.7 79.7%
  18. 18 DeepSeek V4 Pro 0813 79.5%
  19. 19 Grok 4.5 79.2%
  20. 20 Gemini 2.5 Pro Jun 2025 79%
  21. 21 Claude Opus 4.8 79%
  22. 22 Claude Sonnet 5.5 78.5%
  23. 23 Claude Sonnet 5 78.3%
  24. 24 Claude Opus 4.1 78%
  25. 25 GLM-5.3 77.8%
  26. 26 Qwen3.5-122B-A10B int4, local 2x Spark 77.3%
  27. 27 GPT-5.1 77%
  28. 28 Gemini 3.6 Flash 76.8%
  29. 29 Llama-3.1-Nemotron-70B-Instruct 76.8%
  30. 30 Gemma 4 31B IT 76.5%
  31. 31 Mistral Large 3 76.2%
  32. 32 Grok 4.6 76.2%
  33. 33 MiniMax-M3 75.8%
  34. 34 GPT-5.4 75.7%
  35. 35 Qwen 3.6 35B-A3B FP8, local Spark 75.7%
  36. 36 Gemini 3.5 Flash 75.7%
  37. 37 Claude Opus 4.5 75.3%
  38. 38 DeepSeek V4 Flash 0731 75.2%
  39. 39 Gemini 2.5 Flash Sep 2025 75%
  40. 40 GPT-5.4 Mini 75%
  41. 41 Llama 3.3 70B 74.7%
  42. 42 Claude 3.7 Sonnet Prior evaluation 74.7%
  43. 43 Claude Sonnet 4.5 74.7%
  44. 44 Llama 4 Scout 74.3%
  45. 45 GPT-5.2 74.3%
  46. 46 Qwen3-Coder-Next FP8, local Spark 73.7%
  47. 47 Solar Pro 4 73.5%
  48. 48 GPT-4o mini Prior evaluation 73.3%
  49. 49 GPT-4o Nov 2024, Prior evaluation 73.3%
  50. 50 Gemma 4 26B A4B local Spark 73.3%
  51. 51 Claude Sonnet 4.6 72.7%
  52. 52 Claude Haiku 4.5 72.3%
  53. 53 MiniMax-M2.7 AWQ, local 2x Spark 72.3%
  54. 54 Qwen3.5-35B-A3B FP8, local Spark 72%
  55. 55 Claude 3.5 Sonnet (October 2024) Prior evaluation 71.3%
  56. 56 gpt-oss-120b MXFP4, local Spark 71.3%
  57. 57 Nemotron 3 Super NVFP4, local Spark 71.3%
  58. 58 Qwen3.6 27B 71.2%
  59. 59 Llama 3-70B Groq, Prior evaluation 70.7%
  60. 60 Qwen3-32B 70.7%
  61. 61 Gemini 2.5 Flash-Lite Sep 2025 70.1%
  62. 62 DiffusionGemma 26B A4B NVFP4, local Spark 69.3%
  63. 63 Claude 3 Haiku Prior evaluation 68.3%
  64. 64 Nemotron 3 Nano 30B-A3B local Spark 68%
  65. 65 Qwen 3.8 27B 67.1%
  66. 66 Claude 3.5 Haiku Prior evaluation 66.7%
  67. 67 GPT-4 Turbo (Jan 2024) Prior evaluation 66%
  68. 68 Llama 3-8B Groq, Prior evaluation 66%
  69. 69 gpt-oss-safeguard-20b 65.3%
  70. 70 GPT-3.5 Turbo Jan 2024, Prior evaluation 64%
  71. 71 gpt-oss-20b 63.7%
  72. 72 Gemma 2 9B Prior evaluation 62.3%
  73. 73 QwQ-32B Prior evaluation 58.7%
  74. 74 Llama 3.1-8B 58%

Source: Official Pre-Flight leaderboard

#EntryScoreSettingDateSource
1 GPT-5.6 Sol 84% Jul 2026 Official leaderboard
2 GPT-5.6 Terra 83.7% Jul 2026 Official leaderboard
3 Kimi K3 83.6% Jul 2026 Official leaderboard
4 GPT-6 Astra 83.6% Sep 2026 Official leaderboard
5 Claude Opus 5.5 82.9% Sep 2026 Official leaderboard
6 GPT-5.5 82.7% Apr 2026 Official leaderboard
7 Gemini 3.8 Flash 82.5% Sep 2026 Official leaderboard
8 Gemini 3.7 Flash 82.2% Aug 2026 Official leaderboard
9 Claude Fable 5 82% Jun 2026 Official leaderboard
10 Claude Fable 5.1 81.9% Sep 2026 Official leaderboard
11 GPT-5.6 Luna 81.3% Jul 2026 Official leaderboard
12 Qwen 3.8 Max 81.2% Aug 2026 Official leaderboard
13 Claude Opus 5 80.5% Jul 2026 Official leaderboard
14 Gemini 4 Argon 80.5% Sep 2026 Official leaderboard
15 GPT-5 80.3% Aug 2025 Official leaderboard
16 Claude Opus 4.6 80.3% Feb 2026 Official leaderboard
17 Claude Opus 4.7 79.7% Apr 2026 Official leaderboard
18 DeepSeek V4 Pro 0813 79.5% Aug 2026 Official leaderboard
19 Grok 4.5 79.2% Jul 2026 Official leaderboard
20 Gemini 2.5 Pro Jun 2025 79% Jun 2025 Official leaderboard
21 Claude Opus 4.8 79% May 2026 Official leaderboard
22 Claude Sonnet 5.5 78.5% Sep 2026 Official leaderboard
23 Claude Sonnet 5 78.3% Jun 2026 Official leaderboard
24 Claude Opus 4.1 78% Aug 2025 Official leaderboard
25 GLM-5.3 77.8% Aug 2026 Official leaderboard
26 Qwen3.5-122B-A10B 77.3% int4, local 2x Spark Feb 2026 Official leaderboard
27 GPT-5.1 77% Nov 2025 Official leaderboard
28 Gemini 3.6 Flash 76.8% Jul 2026 Official leaderboard
29 Llama-3.1-Nemotron-70B-Instruct 76.8% Oct 2024 Official leaderboard
30 Gemma 4 31B IT 76.5% Apr 2026 Official leaderboard
31 Mistral Large 3 76.2% Dec 2025 Official leaderboard
32 Grok 4.6 76.2% Aug 2026 Official leaderboard
33 MiniMax-M3 75.8% Jun 2026 Official leaderboard
34 GPT-5.4 75.7% Mar 2026 Official leaderboard
35 Qwen 3.6 35B-A3B 75.7% FP8, local Spark Apr 2026 Official leaderboard
36 Gemini 3.5 Flash 75.7% May 2026 Official leaderboard
37 Claude Opus 4.5 75.3% Nov 2025 Official leaderboard
38 DeepSeek V4 Flash 0731 75.2% Jul 2026 Official leaderboard
39 Gemini 2.5 Flash Sep 2025 75% Sep 2025 Official leaderboard
40 GPT-5.4 Mini 75% Mar 2026 Official leaderboard
41 Llama 3.3 70B 74.7% Dec 2024 Official leaderboard
42 Claude 3.7 Sonnet 74.7% Prior evaluation Feb 2025 Official leaderboard
43 Claude Sonnet 4.5 74.7% Sep 2025 Official leaderboard
44 Llama 4 Scout 74.3% Apr 2025 Official leaderboard
45 GPT-5.2 74.3% Dec 2025 Official leaderboard
46 Qwen3-Coder-Next 73.7% FP8, local Spark Feb 2026 Official leaderboard
47 Solar Pro 4 73.5% Aug 2026 Official leaderboard
48 GPT-4o mini 73.3% Prior evaluation Jul 2024 Official leaderboard
49 GPT-4o Nov 2024 73.3% Prior evaluation Nov 2024 Official leaderboard
50 Gemma 4 26B A4B 73.3% local Spark Apr 2026 Official leaderboard
51 Claude Sonnet 4.6 72.7% Feb 2026 Official leaderboard
52 Claude Haiku 4.5 72.3% Oct 2025 Official leaderboard
53 MiniMax-M2.7 72.3% AWQ, local 2x Spark Mar 2026 Official leaderboard
54 Qwen3.5-35B-A3B 72% FP8, local Spark Feb 2026 Official leaderboard
55 Claude 3.5 Sonnet (October 2024) 71.3% Prior evaluation Oct 2024 Official leaderboard
56 gpt-oss-120b 71.3% MXFP4, local Spark Aug 2025 Official leaderboard
57 Nemotron 3 Super 71.3% NVFP4, local Spark Mar 2026 Official leaderboard
58 Qwen3.6 27B 71.2% Apr 2026 Official leaderboard
59 Llama 3-70B 70.7% Groq, Prior evaluation Apr 2024 Official leaderboard
60 Qwen3-32B 70.7% Apr 2025 Official leaderboard
61 Gemini 2.5 Flash-Lite Sep 2025 70.1% Sep 2025 Official leaderboard
62 DiffusionGemma 26B A4B 69.3% NVFP4, local Spark Jun 2026 Official leaderboard
63 Claude 3 Haiku 68.3% Prior evaluation Mar 2024 Official leaderboard
64 Nemotron 3 Nano 30B-A3B 68% local Spark Dec 2025 Official leaderboard
65 Qwen 3.8 27B 67.1% Aug 2026 Official leaderboard
66 Claude 3.5 Haiku 66.7% Prior evaluation Oct 2024 Official leaderboard
67 GPT-4 Turbo (Jan 2024) 66% Prior evaluation Jan 2024 Official leaderboard
68 Llama 3-8B 66% Groq, Prior evaluation Apr 2024 Official leaderboard
69 gpt-oss-safeguard-20b 65.3% Oct 2025 Official leaderboard
70 GPT-3.5 Turbo Jan 2024 64% Prior evaluation Jan 2024 Official leaderboard
71 gpt-oss-20b 63.7% Aug 2025 Official leaderboard
72 Gemma 2 9B 62.3% Prior evaluation Jun 2024 Official leaderboard
73 QwQ-32B 58.7% Prior evaluation Mar 2025 Official leaderboard
74 Llama 3.1-8B 58% Jul 2024 Official leaderboard
75 ALLaM 2 7B 25.3% Official leaderboard
#ModelScoreSettingDateSource
1 GPT-5.6 Sol 84% Jul 2026 Official leaderboard
2 GPT-5.6 Terra 83.7% Jul 2026 Official leaderboard
3 Kimi K3 83.6% Jul 2026 Official leaderboard
4 GPT-6 Astra 83.6% Sep 2026 Official leaderboard
5 Claude Opus 5.5 82.9% Sep 2026 Official leaderboard
6 GPT-5.5 82.7% Apr 2026 Official leaderboard
7 Gemini 3.8 Flash 82.5% Sep 2026 Official leaderboard
8 Gemini 3.7 Flash 82.2% Aug 2026 Official leaderboard
9 Claude Fable 5 82% Jun 2026 Official leaderboard
10 Claude Fable 5.1 81.9% Sep 2026 Official leaderboard
11 GPT-5.6 Luna 81.3% Jul 2026 Official leaderboard
12 Qwen 3.8 Max 81.2% Aug 2026 Official leaderboard
13 Claude Opus 5 80.5% Jul 2026 Official leaderboard
14 Gemini 4 Argon 80.5% Sep 2026 Official leaderboard
15 GPT-5 80.3% Aug 2025 Official leaderboard
16 Claude Opus 4.6 80.3% Feb 2026 Official leaderboard
17 Claude Opus 4.7 79.7% Apr 2026 Official leaderboard
18 DeepSeek V4 Pro 0813 79.5% Aug 2026 Official leaderboard
19 Grok 4.5 79.2% Jul 2026 Official leaderboard
20 Gemini 2.5 Pro Jun 2025 79% Jun 2025 Official leaderboard
21 Claude Opus 4.8 79% May 2026 Official leaderboard
22 Claude Sonnet 5.5 78.5% Sep 2026 Official leaderboard
23 Claude Sonnet 5 78.3% Jun 2026 Official leaderboard
24 Claude Opus 4.1 78% Aug 2025 Official leaderboard
25 GLM-5.3 77.8% Aug 2026 Official leaderboard
26 Qwen3.5-122B-A10B 77.3% int4, local 2x Spark Feb 2026 Official leaderboard
27 GPT-5.1 77% Nov 2025 Official leaderboard
28 Gemini 3.6 Flash 76.8% Jul 2026 Official leaderboard
29 Llama-3.1-Nemotron-70B-Instruct 76.8% Oct 2024 Official leaderboard
30 Gemma 4 31B IT 76.5% Apr 2026 Official leaderboard
31 Mistral Large 3 76.2% Dec 2025 Official leaderboard
32 Grok 4.6 76.2% Aug 2026 Official leaderboard
33 MiniMax-M3 75.8% Jun 2026 Official leaderboard
34 GPT-5.4 75.7% Mar 2026 Official leaderboard
35 Qwen 3.6 35B-A3B 75.7% FP8, local Spark Apr 2026 Official leaderboard
36 Gemini 3.5 Flash 75.7% May 2026 Official leaderboard
37 Claude Opus 4.5 75.3% Nov 2025 Official leaderboard
38 DeepSeek V4 Flash 0731 75.2% Jul 2026 Official leaderboard
39 Gemini 2.5 Flash Sep 2025 75% Sep 2025 Official leaderboard
40 GPT-5.4 Mini 75% Mar 2026 Official leaderboard
41 Llama 3.3 70B 74.7% Dec 2024 Official leaderboard
42 Claude 3.7 Sonnet 74.7% Prior evaluation Feb 2025 Official leaderboard
43 Claude Sonnet 4.5 74.7% Sep 2025 Official leaderboard
44 Llama 4 Scout 74.3% Apr 2025 Official leaderboard
45 GPT-5.2 74.3% Dec 2025 Official leaderboard
46 Qwen3-Coder-Next 73.7% FP8, local Spark Feb 2026 Official leaderboard
47 Solar Pro 4 73.5% Aug 2026 Official leaderboard
48 GPT-4o mini 73.3% Prior evaluation Jul 2024 Official leaderboard
49 GPT-4o Nov 2024 73.3% Prior evaluation Nov 2024 Official leaderboard
50 Gemma 4 26B A4B 73.3% local Spark Apr 2026 Official leaderboard
51 Claude Sonnet 4.6 72.7% Feb 2026 Official leaderboard
52 Claude Haiku 4.5 72.3% Oct 2025 Official leaderboard
53 MiniMax-M2.7 72.3% AWQ, local 2x Spark Mar 2026 Official leaderboard
54 Qwen3.5-35B-A3B 72% FP8, local Spark Feb 2026 Official leaderboard
55 Claude 3.5 Sonnet (October 2024) 71.3% Prior evaluation Oct 2024 Official leaderboard
56 gpt-oss-120b 71.3% MXFP4, local Spark Aug 2025 Official leaderboard
57 Nemotron 3 Super 71.3% NVFP4, local Spark Mar 2026 Official leaderboard
58 Qwen3.6 27B 71.2% Apr 2026 Official leaderboard
59 Llama 3-70B 70.7% Groq, Prior evaluation Apr 2024 Official leaderboard
60 Qwen3-32B 70.7% Apr 2025 Official leaderboard
61 Gemini 2.5 Flash-Lite Sep 2025 70.1% Sep 2025 Official leaderboard
62 DiffusionGemma 26B A4B 69.3% NVFP4, local Spark Jun 2026 Official leaderboard
63 Claude 3 Haiku 68.3% Prior evaluation Mar 2024 Official leaderboard
64 Nemotron 3 Nano 30B-A3B 68% local Spark Dec 2025 Official leaderboard
65 Qwen 3.8 27B 67.1% Aug 2026 Official leaderboard
66 Claude 3.5 Haiku 66.7% Prior evaluation Oct 2024 Official leaderboard
67 GPT-4 Turbo (Jan 2024) 66% Prior evaluation Jan 2024 Official leaderboard
68 Llama 3-8B 66% Groq, Prior evaluation Apr 2024 Official leaderboard
69 gpt-oss-safeguard-20b 65.3% Oct 2025 Official leaderboard
70 GPT-3.5 Turbo Jan 2024 64% Prior evaluation Jan 2024 Official leaderboard
71 gpt-oss-20b 63.7% Aug 2025 Official leaderboard
72 Gemma 2 9B 62.3% Prior evaluation Jun 2024 Official leaderboard
73 QwQ-32B 58.7% Prior evaluation Mar 2025 Official leaderboard
74 Llama 3.1-8B 58% Jul 2024 Official leaderboard

Scores at each effort level

Models the leaderboard ran at two or more effort levels. We don’t have enough cost data here to say what the extra effort costs.

Lowest effort to highest, light to dark
Llama 4 Scout
Prior evalu…
74.3%
Qwen3-Coder-Next
int4, local… FP8, local …
73.7%
gpt-oss-120b
MXFP4, loca…
71.3%

Source: Official Pre-Flight leaderboard

ModelEffortHarnessScoreCost per taskOutput tokens
GPT-5.6 Sol 84%
GPT-5.6 Terra 83.7%
Kimi K3 83.6%
GPT-6 Astra 83.6%
Claude Opus 5.5 82.9%
GPT-5.5 82.7%
Gemini 3.8 Flash 82.5%
Gemini 3.7 Flash 82.2%
Claude Fable 5 82%
Claude Fable 5.1 81.9%
GPT-5.6 Luna 81.3%
Qwen 3.8 Max 81.2%
Claude Opus 5 80.5%
Gemini 4 Argon 80.5%
GPT-5 80.3%
Claude Opus 4.6 80.3%
Claude Opus 4.7 79.7%
DeepSeek V4 Pro 0813 79.5%
Grok 4.5 79.2%
Gemini 2.5 Pro Jun 2025 79%
Claude Opus 4.8 79%
Claude Sonnet 5.5 78.5%
Claude Sonnet 5 78.3%
Claude Opus 4.1 78%
GLM-5.3 77.8%
Qwen3.5-122B-A10B int4, local 2x Spark 77.3%
GPT-5.1 77%
Gemini 3.6 Flash 76.8%
Llama-3.1-Nemotron-70B-Instruct 76.8%
Gemma 4 31B IT 76.5%
Mistral Large 3 76.2%
Grok 4.6 76.2%
MiniMax-M3 75.8%
GPT-5.4 75.7%
Qwen 3.6 35B-A3B FP8, local Spark 75.7%
Gemini 3.5 Flash 75.7%
Claude Opus 4.5 75.3%
DeepSeek V4 Flash 0731 75.2%
Gemini 2.5 Flash Sep 2025 75%
GPT-5.4 Mini 75%
Llama 3.3 70B 74.7%
Claude 3.7 Sonnet Prior evaluation 74.7%
Claude Sonnet 4.5 74.7%
Llama 4 Scout 74.3%
GPT-5.2 74.3%
Qwen3-Coder-Next FP8, local Spark 73.7%
Solar Pro 4 73.5%
GPT-4o mini Prior evaluation 73.3%
GPT-4o Nov 2024 Prior evaluation 73.3%
Gemma 4 26B A4B local Spark 73.3%
Claude Sonnet 4.6 72.7%
Claude Haiku 4.5 72.3%
MiniMax-M2.7 AWQ, local 2x Spark 72.3%
Qwen3.5-35B-A3B FP8, local Spark 72%
Llama 4 Scout Prior evaluation 71.7%
Claude 3.5 Sonnet (October 2024) Prior evaluation 71.3%
gpt-oss-120b MXFP4, local Spark 71.3%
Qwen3-Coder-Next int4, local Spark 71.3%
Nemotron 3 Super NVFP4, local Spark 71.3%
Qwen3.6 27B 71.2%
Llama 3-70B Groq, Prior evaluation 70.7%
Qwen3-32B 70.7%
gpt-oss-120b 70.7%
Gemini 2.5 Flash-Lite Sep 2025 70.1%
DiffusionGemma 26B A4B NVFP4, local Spark 69.3%
Claude 3 Haiku Prior evaluation 68.3%
Nemotron 3 Nano 30B-A3B local Spark 68%
Qwen 3.8 27B 67.1%
Claude 3.5 Haiku Prior evaluation 66.7%
GPT-4 Turbo (Jan 2024) Prior evaluation 66%
Llama 3-8B Groq, Prior evaluation 66%
gpt-oss-safeguard-20b 65.3%
GPT-3.5 Turbo Jan 2024 Prior evaluation 64%
gpt-oss-20b 63.7%
Gemma 2 9B Prior evaluation 62.3%
QwQ-32B Prior evaluation 58.7%
Llama 3.1-8B 58%

Scores at each effort level

Models the leaderboard ran at two or more effort levels. We don’t have enough cost data here to say what the extra effort costs.

Lowest effort to highest, light to dark
Llama 4 Scout
Prior evalu…
74.3%
Qwen3-Coder-Next
int4, local… FP8, local …
73.7%
gpt-oss-120b
MXFP4, loca…
71.3%

Source: Official Pre-Flight leaderboard

ModelEffortHarnessScoreCost per taskOutput tokens
GPT-5.6 Sol 84%
GPT-5.6 Terra 83.7%
Kimi K3 83.6%
GPT-6 Astra 83.6%
Claude Opus 5.5 82.9%
GPT-5.5 82.7%
Gemini 3.8 Flash 82.5%
Gemini 3.7 Flash 82.2%
Claude Fable 5 82%
Claude Fable 5.1 81.9%
GPT-5.6 Luna 81.3%
Qwen 3.8 Max 81.2%
Claude Opus 5 80.5%
Gemini 4 Argon 80.5%
GPT-5 80.3%
Claude Opus 4.6 80.3%
Claude Opus 4.7 79.7%
DeepSeek V4 Pro 0813 79.5%
Grok 4.5 79.2%
Gemini 2.5 Pro Jun 2025 79%
Claude Opus 4.8 79%
Claude Sonnet 5.5 78.5%
Claude Sonnet 5 78.3%
Claude Opus 4.1 78%
GLM-5.3 77.8%
Qwen3.5-122B-A10B int4, local 2x Spark 77.3%
GPT-5.1 77%
Gemini 3.6 Flash 76.8%
Llama-3.1-Nemotron-70B-Instruct 76.8%
Gemma 4 31B IT 76.5%
Mistral Large 3 76.2%
Grok 4.6 76.2%
MiniMax-M3 75.8%
GPT-5.4 75.7%
Qwen 3.6 35B-A3B FP8, local Spark 75.7%
Gemini 3.5 Flash 75.7%
Claude Opus 4.5 75.3%
DeepSeek V4 Flash 0731 75.2%
Gemini 2.5 Flash Sep 2025 75%
GPT-5.4 Mini 75%
Llama 3.3 70B 74.7%
Claude 3.7 Sonnet Prior evaluation 74.7%
Claude Sonnet 4.5 74.7%
Llama 4 Scout 74.3%
GPT-5.2 74.3%
Qwen3-Coder-Next FP8, local Spark 73.7%
Solar Pro 4 73.5%
GPT-4o mini Prior evaluation 73.3%
GPT-4o Nov 2024 Prior evaluation 73.3%
Gemma 4 26B A4B local Spark 73.3%
Claude Sonnet 4.6 72.7%
Claude Haiku 4.5 72.3%
MiniMax-M2.7 AWQ, local 2x Spark 72.3%
Qwen3.5-35B-A3B FP8, local Spark 72%
Llama 4 Scout Prior evaluation 71.7%
Claude 3.5 Sonnet (October 2024) Prior evaluation 71.3%
gpt-oss-120b MXFP4, local Spark 71.3%
Qwen3-Coder-Next int4, local Spark 71.3%
Nemotron 3 Super NVFP4, local Spark 71.3%
Qwen3.6 27B 71.2%
Llama 3-70B Groq, Prior evaluation 70.7%
Qwen3-32B 70.7%
gpt-oss-120b 70.7%
Gemini 2.5 Flash-Lite Sep 2025 70.1%
DiffusionGemma 26B A4B NVFP4, local Spark 69.3%
Claude 3 Haiku Prior evaluation 68.3%
Nemotron 3 Nano 30B-A3B local Spark 68%
Qwen 3.8 27B 67.1%
Claude 3.5 Haiku Prior evaluation 66.7%
GPT-4 Turbo (Jan 2024) Prior evaluation 66%
Llama 3-8B Groq, Prior evaluation 66%
gpt-oss-safeguard-20b 65.3%
GPT-3.5 Turbo Jan 2024 Prior evaluation 64%
gpt-oss-20b 63.7%
Gemma 2 9B Prior evaluation 62.3%
QwQ-32B Prior evaluation 58.7%
Llama 3.1-8B 58%

5 models have held number one on the official leaderboard since its release in March 2025

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

NVIDIA Google DeepMind OpenAI
Own leaderboard Not tested since Sep 2026

Source: Official Pre-Flight 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 the benchmark’s own leaderboard Best so far
0% 20% 40% 60% 80% 100% 2024 2025 2026 Near the ceiling (85%) GPT-5.6 Sol GPT-5 Llama-3.1-Nemotron-70B-Instruct Llama 3-70B GPT-4 Turbo (Jan 2024) 0% 20% 40% 60% 80% 100% 2024 2025 2026 Near the ceiling (85%) GPT-5.6 Sol Llama-3.1-Nemotron-70B-Instruct

Source: Official Pre-Flight 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 Law, finance and medicine

All 637 in Law, finance and medicine
  • MedXpertQA-MM In use Models answer expert-level medical exam questions that combine images with clinical records and test results. Scoring is accuracy. Qwen 3.8 Max 80.4%
  • HealthBench Professional In use 525 physician-authored tasks from real clinician chats (care consult, documentation, research) with physician-adjudicated rubrics. Claude Sonnet 5.5 69.2%
  • HealthBench In use 5,000 realistic health conversations graded against physician-written rubrics. A length-adjusted variant penalizes verbosity. GPT-5 67.2%
  • FinSearchComp In use Agents search the web to fetch live financial data, look up history, and run multi-step investigations, across 635 expert-written questions. Scored by answer accuracy. Ling-3.1-flash 81.7%
  • KorMedMCQA In use Models answer 7,469 multiple-choice questions from Korean licensing exams for doctors, nurses, pharmacists and dentists, held from 2012 to 2024. o1-preview 92.7%
  • Harvey LAB-AA v1.1 In use Artificial Analysis's run of Harvey's Legal Agent Benchmark on Harvey's newer private set of 120 tasks across 24 practice areas. A deliverable with a material hallucination scores zero, and the headline is the share of judges who find every rubric criterion passed. GPT-6.2 Astra 15%

Data audit

All 11 checks pass. Every benchmark

Catalog record 2 pass
  • Pass Task access All tasks are public.
  • Pass Confidence Checked against a primary source.
Scores 5 pass
  • Pass Independent runs Scores from the official leaderboard (75 entries).
  • Pass Models scored 74 models we track, plus 1 other entry on its leaderboard.
  • Pass Newest score The newest is from September 2026, for Gemini 4 Argon.
  • Pass Frontier models 5 of the top 10 have a score, led by GPT-6 Astra (#4).
  • Pass Headroom The best is 84%, under 85%, so it still separates models.
Leaderboard import 3 pass
  • Pass Importer We read the board on airsidelabs.com: 77 rows for models we track and 1 other entry 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