Results
| Position | |||||||||
|---|---|---|---|---|---|---|---|---|---|
| 1 | Gemini 4 Argon (multi-step) Google, low-thinking draft, medium-thinking checks Character error 1.5%, colour 98, microfilm 99, 7 of 7 run, 4.5 min | 98.1 | 1.5% | 5.9% | 98 | 99 | 7 | $0.040 | 4.5 min |
| 2 | Gemini 4 Argon (high) Google, high thinking Character error 1.6%, colour 97, microfilm 99, 7 of 7 run, 12 min | 97.9 | 1.6% | 6.2% | 97 | 99 | 7 | $0.080 | 12 min |
| 3 | Gemini 4 Argon (low) Google, low thinking Character error 2.2%, colour 97, microfilm 98, 7 of 7 run, 1.9 min | 97.3 | 2.2% | 7.5% | 97 | 98 | 7 | $0.0089 | 1.9 min |
| =4 | Gemini 3 Pro (high) Google, high thinking Character error 4.6%, colour 95, microfilm 94, 7 of 7 run, 16 min | 94.9 | 4.6% | 14.3% | 95 | 94 | 7 | $0.17 | 16 min |
| =4 | Gemini 3.8 Flash (low) Google, low thinking Character error 4.6%, colour 95, microfilm 94, 7 of 7 run, 76 s | 94.9 | 4.6% | 13.9% | 95 | 94 | 7 | $0.0035 | 76 s |
| =4 | Gemini 3 Flash (low) Google, low thinking Character error 4.8%, colour 96, microfilm 93, 7 of 7 run, 50 s | 94.9 | 4.8% | 13.8% | 96 | 93 | 7 | $0.0027 | 50 s |
| 7 | Opus 5.5 (high) Anthropic, high thinking Character error 5.0%, colour 96, microfilm 92, 7 of 7 run, 4.8 min | 94.6 | 5.0% | 15.5% | 96 | 92 | 7 | $0.11 | 4.8 min |
| 8 | Gemini 3.5 Flash (low) Google, low thinking Character error 5.1%, colour 95, microfilm 94, 7 of 7 run, 1.8 min | 94.4 | 5.1% | 14.6% | 95 | 94 | 7 | $0.0024 | 1.8 min |
| =9 | Gemini 3.5 Flash (high) Google, high thinking Character error 5.9%, colour 95, microfilm 90, 7 of 7 run, 9.6 min | 93.6 | 5.9% | 15.2% | 95 | 90 | 7 | – | 9.6 min |
| =9 | Gemini 3 Pro (low) Google, low thinking Character error 5.9%, colour 95, microfilm 89, 7 of 7 run, 1.8 min | 93.6 | 5.9% | 15.7% | 95 | 89 | 7 | $0.013 | 1.8 min |
| 11 | Gemini 3 Flash (high) Google, high thinking Character error 6.8%, colour 94, microfilm 89, 7 of 7 run, 9.7 min | 92.6 | 6.8% | 19.2% | 94 | 89 | 7 | $0.042 | 9.7 min |
| 12 | Gemini 3.1 Pro (low) Google, low thinking Character error 7.5%, colour 92, microfilm 92, 7 of 7 run, 1.9 min | 92.0 | 7.5% | 26.0% | 92 | 92 | 7 | $0.011 | 1.9 min |
| 13 | Gemini 3.1 Pro (high) Google, high thinking Character error 9.0%, colour 92, microfilm 88, 7 of 7 run, 20 min | 90.6 | 9.0% | 25.4% | 92 | 88 | 7 | $0.19 | 20 min |
| 14 | Opus 5.5 (low) Anthropic, low thinking Character error 9.4%, colour 95, microfilm 77, 7 of 7 run, 1.5 min | 90.1 | 9.4% | 21.0% | 95 | 77 | 7 | $0.037 | 1.5 min |
| 15 | GPT-6.1 Sol (high) OpenAI, high thinking Character error 21.1%, colour 96, microfilm 9, 5 of 7 run, 1 h 23 min | 78.2 | 21.1% | 27.2% | 96 | 9 | 5 of 72 failed | $0.11 | 1 h 23 min |
| 16 | Sonnet 5.5 (low) Anthropic, low thinking Character error 24.9%, colour 93, microfilm 30, 7 of 7 run, 86 s | 74.6 | 24.9% | 56.6% | 93 | 30 | 7 | $0.019 | 86 s |
| 17 | Opus 4.6 (thinking) Anthropic, thinking on Character error 27.8%, colour 85, microfilm 40, 7 of 7 run, 6.9 min | 71.9 | 27.8% | 93.5% | 85 | 40 | 7 | $0.037 | 6.9 min |
| 18 | Sonnet 5.5 (high) Anthropic, high thinking Character error 28.0%, colour 86, microfilm 37, 7 of 7 run, 74 s | 71.6 | 28.0% | 42.8% | 86 | 37 | 7 | $0.017 | 74 s |
| 19 | Opus 4.7 Anthropic, default settings Character error 28.3%, colour 88, microfilm 29, 7 of 7 run, 5.8 min | 71.3 | 28.3% | 56.5% | 88 | 29 | 7 | $0.053 | 5.8 min |
| 20 | Muse Spark Open OpenRouter, default settings Character error 28.7%, colour 85, microfilm 36, 7 of 7 run, 4.7 min | 70.9 | 28.7% | 54.5% | 85 | 36 | 7 | $0.021 | 4.7 min |
| 21 | Opus 4.7 (thinking) Anthropic, thinking on Character error 29.5%, colour 88, microfilm 26, 7 of 7 run, 5.5 min | 70.1 | 29.5% | 58.5% | 88 | 26 | 7 | $0.052 | 5.5 min |
| 22 | Opus 4.6 Anthropic, default settings Character error 29.9%, colour 83, microfilm 37, 7 of 7 run, 5.8 min | 69.7 | 29.9% | 89.3% | 83 | 37 | 7 | $0.036 | 5.8 min |
| 23 | GPT-6.1 Sol (low) OpenAI, low thinking Character error 30.7%, colour 89, microfilm 19, 7 of 7 run, 3.4 min | 68.9 | 30.7% | 52.5% | 89 | 19 | 7 | $0.044 | 3.4 min |
| 24 | Sonnet 4.6 Anthropic, default settings Character error 34.2%, colour 79, microfilm 31, 7 of 7 run, 5.5 min | 65.4 | 34.2% | 86.5% | 79 | 31 | 7 | $0.022 | 5.5 min |
| 25 | Kimi K3 (no thinking) OpenRouter, thinking off Character error 34.5%, colour 84, microfilm 18, 7 of 7 run, 2.1 min | 65.0 | 34.5% | 59.8% | 84 | 18 | 7 | $0.026 | 2.1 min |
| 26 | Kimi K3 OpenRouter, default settings Character error 35.8%, colour 84, microfilm 24, 6 of 7 run, 17 min | 63.7 | 35.8% | 60.9% | 84 | 24 | 6 of 71 failed | $0.089 | 17 min |
| 27 | Kimi K2.6 OpenRouter, default settings Character error 36.8%, colour 80, microfilm 28, 6 of 7 run, 53 min | 62.7 | 36.8% | 61.4% | 80 | 28 | 6 of 71 failed | $0.11 | 53 min |
| 28 | Haiku 5.5 (low) Anthropic, low thinking Character error 38.3%, colour 74, microfilm 30, 7 of 7 run, 64 s | 61.4 | 38.3% | 62.9% | 74 | 30 | 7 | $0.0011 | 64 s |
| =29 | Haiku 5.5 (high) Anthropic, high thinking Character error 38.4%, colour 77, microfilm 22, 7 of 7 run, 2.4 min | 61.0 | 38.4% | 63.1% | 77 | 22 | 7 | $0.0027 | 2.4 min |
| =29 | Sonnet 4.6 (thinking) Anthropic, thinking on Character error 38.5%, colour 73, microfilm 30, 7 of 7 run, 7.3 min | 61.0 | 38.5% | 118.9% | 73 | 30 | 7 | $0.051 | 7.3 min |
| 31 | DeepSeek V4.1 Flash 1017 OpenRouter, default settings Character error 41.5%, colour 70, microfilm 12, 5 of 7 run, 12 min | 58.0 | 41.5% | 62.1% | 70 | 12 | 5 of 72 failed | $0.0047 | 12 min |
| 32 | Gemini 3.8 Flash (high) Google, high thinking Character error 41.7%, colour 63, microfilm 45, 7 of 7 run, 15 min | 57.6 | 41.7% | 57.9% | 63 | 45 | 7 | $0.061 | 15 min |
| 33 | GPT-6 Astra (low) OpenAI, low thinking Character error 62.2%, colour 77, microfilm 0, 7 of 7 run, 4.6 min | 55.0 | 62.2% | 85.7% | 77 | 0 | 7 | $0.23 | 4.6 min |
| 34 | Qwen 3.6 Plus OpenRouter, default settings Character error 54.3%, colour 66, microfilm 26, 7 of 7 run, 13 min | 54.4 | 54.3% | 159.0% | 66 | 26 | 7 | $0.022 | 13 min |
| 35 | GPT-5.2 (low) OpenAI, low thinking Character error 46.9%, colour 74, microfilm 0, 7 of 7 run, 2.8 min | 52.9 | 46.9% | 64.9% | 74 | 0 | 7 | $0.018 | 2.8 min |
| 36 | GPT-5.2 (high) OpenAI, high thinking Character error 48.1%, colour 70, microfilm 6, 7 of 7 run, 11 min | 51.3 | 48.1% | 72.6% | 70 | 6 | 7 | $0.075 | 11 min |
| 37 | GPT-5.5 (low) OpenAI, low thinking Character error 53.2%, colour 49, microfilm 40, 7 of 7 run, 5 min | 46.4 | 53.2% | 74.8% | 49 | 40 | 7 | $0.065 | 5 min |
| 38 | DeepSeek V4.1 Flash 1017 (no thinking) OpenRouter, thinking off Character error 61.1%, colour 59, microfilm 13, 7 of 7 run, 78 s | 45.6 | 61.1% | 84.7% | 59 | 13 | 7 | $0.0008 | 78 s |
| 39 | Qwen 3.6 35B A3B OpenRouter, default settings Character error 55.2%, colour 51, microfilm 27, 7 of 7 run, 7.2 min | 44.3 | 55.2% | 110.7% | 51 | 27 | 7 | $0.0074 | 7.2 min |
| 40 | Grok 4.3 OpenRouter, default settings Character error 75.5%, colour 56, microfilm 6, 7 of 7 run, 4.3 min | 41.3 | 75.5% | 149.7% | 56 | 6 | 7 | $0.0076 | 4.3 min |
| 41 | GPT-4.1 OpenAI, default settings Character error 58.6%, colour 49, microfilm 20, 7 of 7 run, 2 min | 40.9 | 58.6% | 90.8% | 49 | 20 | 7 | $0.0071 | 2 min |
| 42 | GPT-5.4 Mini (low) OpenAI, low thinking Character error 59.0%, colour 51, microfilm 14, 7 of 7 run, 1.6 min | 40.4 | 59.0% | 85.4% | 51 | 14 | 7 | $0.0063 | 1.6 min |
| 43 | Qwen 3.8 27B OpenRouter, default settings Character error 59.4%, colour 46, microfilm 25, 7 of 7 run, 7.8 min | 40.3 | 59.4% | 177.4% | 46 | 25 | 7 | $0.0043 | 7.8 min |
| 44 | GPT-5 Mini (low) OpenAI, low thinking Character error 62.5%, colour 46, microfilm 15, 7 of 7 run, 2 min | 37.0 | 62.5% | 95.4% | 46 | 15 | 7 | $0.0023 | 2 min |
| 45 | Haiku 4.5 Anthropic, default settings Character error 63.0%, colour 44, microfilm 18, 7 of 7 run, 2.3 min | 36.4 | 63.0% | 123.1% | 44 | 18 | 7 | $0.0066 | 2.3 min |
| 46 | GPT-5 Mini (high) OpenAI, high thinking Character error 69.4%, colour 47, microfilm 0, 7 of 7 run, 8.3 min | 33.4 | 69.4% | 108.6% | 47 | 0 | 7 | $0.0084 | 8.3 min |
| 47 | Haiku 4.5 (thinking) Anthropic, thinking on Character error 66.9%, colour 39, microfilm 17, 7 of 7 run, 3.4 min | 32.6 | 66.9% | 142.1% | 39 | 17 | 7 | $0.015 | 3.4 min |
| 48 | Llama 4 Maverick OpenRouter, default settings Character error 82.6%, colour 30, microfilm 28, 7 of 7 run, 4.5 min | 29.4 | 82.6% | 126.8% | 30 | 28 | 7 | $0.0008 | 4.5 min |
| 49 | Llama 4 Scout OpenRouter, default settings Character error 70.4%, colour 33, microfilm 19, 7 of 7 run, 2.6 min | 29.3 | 70.4% | 96.6% | 33 | 19 | 7 | $0.0005 | 2.6 min |
| 50 | Nemotron 3.5 Super VL OpenRouter, default settings Character error 73.9%, colour 25, microfilm –, 3 of 7 run, 17 min | 25.3 | 73.9% | 100.8% | 25 | – | 3 of 74 failed | $0.0031 | 17 min |
| 51 | GLM 5.3 Flash OpenRouter, default settings Character error 75.2%, colour 34, microfilm 1, 7 of 7 run, 2.6 min | 24.3 | 75.2% | 91.0% | 34 | 1 | 7 | $0.0009 | 2.6 min |
| 52 | GPT-5.4 Mini (high) OpenAI, high thinking Character error 76.1%, colour 32, microfilm 2, 7 of 7 run, 12 min | 23.4 | 76.1% | 86.1% | 32 | 2 | 7 | $0.061 | 12 min |
| 53 | Nemotron 3.5 Super VL (no thinking) OpenRouter, thinking off Character error 799.8%, colour 7, microfilm 0, 7 of 7 run, 2.9 min | 5.1 | 799.8% | 1265.8% | 7 | 0 | 7 | $0.0052 | 2.9 min |
| 54 | GPT-6 Luna (low) OpenAI, low thinking Character error 96.7%, colour 4, microfilm 1, 7 of 7 run, 73 s | 3.0 | 96.7% | 98.6% | 4 | 1 | 7 | $0.0017 | 73 s |
| 55 | GPT-6 Luna (high) OpenAI, high thinking Character error 98.7%, colour 0, microfilm 3, 7 of 7 run, 5.7 min | 1.0 | 98.7% | 99.4% | 0 | 3 | 7 | $0.0035 | 5.7 min |
55 of 55 models