AA artificial_analysis_codi   score  Nous route                                   in$/M  out$/M  score/$
Claude Fable 5.1 (Max, Defau   81.6  anthropic/claude-fable-5.1:batch              5.00   25.00      8.2
Claude Fable 5.1 (Xhigh, Def   80.7  anthropic/claude-fable-5.1:batch              5.00   25.00      8.1
Claude Fable 5.1 (High, Defa   79.1  anthropic/claude-fable-5.1:batch              5.00   25.00      7.9
GPT-5.6 Sol (Xhigh)            78.3  openai/gpt-5.6-sol:batch                      1.00    5.00     39.1
Claude Opus 5 (Max)            78.0  anthropic/claude-opus-5:batch                 2.50   12.50     15.6
GPT-5.6 Sol (Max)              77.4  openai/gpt-5.6-sol:batch                      1.00    5.00     38.7
GPT-5.6 Sol (High)             77.2  openai/gpt-5.6-sol:batch                      1.00    5.00     38.6
GPT-6 Astra (High)             77.1  openai/gpt-6-astra:batch                      5.00   25.00      7.7
Claude Fable 5.1 (Medium, De   77.1  anthropic/claude-fable-5.1:batch              5.00   25.00      7.7
Claude Opus 5 (Xhigh)          77.0  anthropic/claude-opus-5:batch                 2.50   12.50     15.4
GPT-6 Astra (Max)              76.9  openai/gpt-6-astra:batch                      5.00   25.00      7.7
Grok 4.6 (High)                76.8  x-ai/grok-4.6                                 2.00    6.00     25.6
GPT-6 Astra (Medium)           76.7  openai/gpt-6-astra:batch                      5.00   25.00      7.7
GPT-5.6 Terra (Max)            76.7  openai/gpt-5.6-terra:batch                    1.00    6.00     34.1
Claude Fable 5 (Max, Opus 4.   76.5  anthropic/claude-fable-5:batch                5.00   25.00      7.7
Claude Opus 5 (High)           76.5  anthropic/claude-opus-5:batch                 2.50   12.50     15.3
Muse Spark 1.3 (Xhigh)         76.5  meta/muse-spark-1.3                           1.25    4.25     38.2
Gemini 3.8 Flash (High)        76.3  google/gemini-3.8-flash:batch                 0.38    1.88    101.7
GPT-5.6 Sol (Medium)           76.3  openai/gpt-5.6-sol:batch                      1.00    5.00     38.1
Qwen3.8 Max (0902)             76.2  qwen/qwen3.8-max-prime                        4.00   12.00     12.7
Kimi K3 (Max)                  76.2  moonshotai/kimi-k3                            0.83   13.00     19.7
Gemini 3.7 Flash (High)        76.1  google/gemini-3.7-flash:batch                 0.38    1.88    101.5
GPT-6 Astra (Xhigh)            75.9  openai/gpt-6-astra:batch                      5.00   25.00      7.6
Grok 4.6 (Xhigh)               75.9  x-ai/grok-4.6                                 2.00    6.00     25.3
Muse Spark 1.3 (Max)           75.8  meta/muse-spark-1.3                           1.25    4.25     37.9
GPT-6 Astra (Low)              75.7  openai/gpt-6-astra:batch                      5.00   25.00      7.6
Claude Fable 5.1 (Low, Defau   75.2  anthropic/claude-fable-5.1:batch              5.00   25.00      7.5
GPT-5.5 (Xhigh)                74.9  openai/gpt-5.5:batch                          2.50   15.00     13.3
GLM-5.3 (Max)                  74.8  z-ai/glm-5.3:batch                            0.45    2.00     89.3
Grok 4.6 (Medium)              74.4  x-ai/grok-4.6                                 2.00    6.00     24.8
Claude Opus 5 (Medium)         74.3  anthropic/claude-opus-5:batch                 2.50   12.50     14.9
Claude Opus 4.8 (Max)          74.3  anthropic/claude-opus-4.8:batch               2.50   12.50     14.9
Gemini 3.8 Flash (Medium)      74.1  google/gemini-3.8-flash:batch                 0.38    1.88     98.8
Claude Opus 4.7 (Max)          73.6  anthropic/claude-opus-4.7:batch               2.50   12.50     14.7
Gemini 3.8 Flash (Low)         73.5  google/gemini-3.8-flash:batch                 0.38    1.88     98.0
Qwen3.8-Flash-Next             73.1  qwen/qwen3.8-flash                            0.15    0.47    317.8
Grok 4.5 (High)                72.4  x-ai/grok-4.5                                 2.00    6.00     24.1
Muse Spark 1.2 (Xhigh)         72.2  meta/muse-spark-1.2                           1.25    4.25     36.1
Kimi K3 (Low)                  72.0  moonshotai/kimi-k3                            0.83   13.00     18.6
Qwen3.8 2.4T A95B              71.9  qwen/qwen3.8-2.4t-a95b                        2.00    6.00     24.0
Qwen3.8 Max                    71.8  (no Nous route)                           
GPT-5.5 (High)                 71.6  openai/gpt-5.5:batch                          2.50   15.00     12.7
GLM 5.3 Flash                  71.5  z-ai/glm-5.3-flash:batch                      0.06    0.20    752.6
Gemini 3.7 Flash (Medium)      71.5  google/gemini-3.7-flash:batch                 0.38    1.88     95.3
Claude Sonnet 5 (Max)          71.5  anthropic/claude-sonnet-5:batch               1.00    5.00     35.8
GPT-5.5 (Medium)               71.5  openai/gpt-5.5:batch                          2.50   15.00     12.7
GPT-5.6 Luna (Max)             71.4  openai/gpt-5.6-luna:batch                     0.10    0.60    317.3
Muse Spark 1.1 (Xhigh)         71.3  meta/muse-spark-1.1                           1.25    4.25     35.6
GPT-5.4 (Xhigh)                71.1  openai/gpt-5.4:batch                          1.25    7.50     25.3
Gemini 3.7 Flash (Low)         71.0  google/gemini-3.7-flash:batch                 0.38    1.88     94.7
GPT-5.6 Terra (Xhigh)          70.6  openai/gpt-5.6-terra:batch                    1.00    6.00     31.4
Gemini 3.5 Flash (High)        70.1  google/gemini-3.5-flash:batch                 0.75    4.50     41.5
GPT-5.6 Sol (Low)              69.7  openai/gpt-5.6-sol:batch                      1.00    5.00     34.9
Gemini 3.6 Flash (High)        69.2  google/gemini-3.6-flash:batch                 0.38    1.88     92.3
DeepSeek V4 Flash 0731 (Max)   69.1  deepseek/deepseek-v4-flash                    0.03    1.28    203.8
DeepSeek V4 Pro 0813 (Max)     68.8  deepseek/deepseek-v4-pro                      0.94    1.89     58.4
GLM-5.2 (Max)                  68.8  z-ai/glm-5.2                                  0.90    2.83     49.8
Gemini 3.1 Pro Preview         68.8  google/gemini-3.1-pro-preview:batch           1.00    6.00     30.6
GPT-5.6 Luna (Xhigh)           68.6  openai/gpt-5.6-luna:batch                     0.10    0.60    304.9
Qwen3.8 27B (Xhigh)            68.1  qwen/qwen3.8-27b                              0.02    4.35     61.6
GPT-5.6 Terra (High)           67.1  openai/gpt-5.6-terra:batch                    1.00    6.00     29.8
Claude Opus 5 (Low)            66.9  anthropic/claude-opus-5:batch                 2.50   12.50     13.4
Claude Sonnet 5 (Non-reasoni   66.4  anthropic/claude-sonnet-5:batch               1.00    5.00     33.2
Grok 4.6 (Low)                 66.3  x-ai/grok-4.6                                 2.00    6.00     22.1
Qwen3.7 Max                    66.0  qwen/qwen3.7-max                              1.48    4.42     29.8
GPT-5.6 Sol (Non-reasoning)    65.1  openai/gpt-5.6-sol:batch                      1.00    5.00     32.5
DeepSeek V4 Flash Vision (Ma   65.0  deepseek/deepseek-v4-flash-vision-exp         0.22    0.65    201.0
GPT-5.6 Terra (Medium)         64.7  openai/gpt-5.6-terra:batch                    1.00    6.00     28.8
Motif 3                        63.5  (no Nous route)                           
GPT-5.6 Luna (High)            63.3  openai/gpt-5.6-luna:batch                     0.10    0.60    281.3
Claude Sonnet 4.6 (Max)        63.0  anthropic/claude-sonnet-4.6:batch             1.50    7.50     21.0
Motif 3 (Beta)                 62.0  (no Nous route)                           
Kimi K2.6 (Reasoning)          61.8  moonshotai/kimi-k2.6                          0.46    2.45     64.3
K2 Horizon 375B A23B           61.5  (no Nous route)                           
Quasar 438B (Max, Based on G   61.2  (no Nous route)                           
GPT-5.5 (Low)                  60.9  openai/gpt-5.5:batch                          2.50   15.00     10.8
Apodex 1.1                     60.8  (no Nous route)                           
Kimi K2.7 Code                 60.8  moonshotai/kimi-k2.7-code                     0.67    3.35     45.3
MiMo-V2.5-Pro (Reasoning)      60.2  xiaomi/mimo-v2.5-pro                          0.30    0.61    158.2
KAT Coder Pro V2               59.5  kwaipilot/kat-coder-pro-v2.5                  0.74    2.96     45.9
DeepSeek V4 Pro 0424 (Max)     59.4  deepseek/deepseek-v4-pro                      0.94    1.89     50.4
Nex-N2-Pro (Based on Qwen3.5   59.1  (no Nous route)                           
Hy3                            58.8  tencent/hy3                                   0.08    0.33    407.3
DeepSeek V4 Pro 0424 (High)    58.7  deepseek/deepseek-v4-pro                      0.94    1.89     49.9
Muse Spark                     58.6  meta/muse-spark-1.3                           1.25    4.25     29.3
MiniMax-M3                     58.6  minimax/minimax-m3                            0.23    0.96    142.1
Qwen3.8 27B (Low)              58.2  qwen/qwen3.8-27b                              0.02    4.35     52.6
GPT-5.6 Terra (Low)            58.1  openai/gpt-5.6-terra:batch                    1.00    6.00     25.8
Ling-3.0-flash-VL              57.0  inclusionai/ling-3.0-flash-vl                 0.02    0.06   1829.9
MiMo-V2.5                      56.8  xiaomi/mimo-v2.5                              0.12    0.24    381.8
GPT-5.5 (Non-reasoning)        56.5  openai/gpt-5.5:batch                          2.50   15.00     10.0
DeepSeek V4 Flash 0420 (Max)   56.2  deepseek/deepseek-v4-flash                    0.03    1.28    165.8
Qwen3.8 27B (Medium)           56.1  qwen/qwen3.8-27b                              0.02    4.35     50.7
GPT-5.4 mini (Xhigh)           56.1  openai/gpt-5.4-mini:batch                     0.38    2.25     66.5
GPT-5.4 nano (Xhigh)           56.1  openai/gpt-5.4-nano:batch                     0.10    0.62    242.6
Qwen3.7 Plus                   55.9  qwen/qwen3.7-plus                             0.32    1.28     99.8
GLM-5.1 (Reasoning)            55.8  z-ai/glm-5.1                                  0.97    3.04     37.6
Ling-3.0-flash-Fin             55.6  inclusionai/ling-3.0-flash-fin:free           0.00    0.00  55600.0
Qwen3.6 Plus                   54.5  qwen/qwen3.6-plus                             0.33    1.95     74.5
Qwen3.6 27B (Reasoning)        53.7  qwen/qwen3.6-27b                              0.30    2.00     74.1
Inkling Small                  52.9  thinkingmachines/inkling-small                0.45    1.20     83.0
Solar Pro 4                    52.7  upstage/solar-pro4                            0.09    0.36    334.6
MiniMax-M2.7                   52.6  minimax/minimax-m2.7                          0.21    0.84    143.1
JT-4.1 Flash 236B A21B (Non-   52.4  (no Nous route)                           
GPT-5.6 Terra (Non-reasoning   52.3  openai/gpt-5.6-terra:batch                    1.00    6.00     23.2
Inkling (Xhigh)                52.1  thinkingmachines/inkling                      0.95    4.05     30.2
Claude 4.5 Sonnet (Reasoning   52.1  (no Nous route)                           
DeepSeek V4 Flash 0420 (High   52.0  deepseek/deepseek-v4-flash                    0.03    1.28    153.4
Grok Build 0.1 0616            51.5  x-ai/grok-build-0.1                           1.00    2.00     41.2
GPT-5.6 Luna (Medium)          50.7  openai/gpt-5.6-luna:batch                     0.10    0.60    225.3
Ling 3.0 Flash                 50.6  inclusionai/ling-3.0-flash                    0.02    0.06   1606.3
MiMo-V2-Flash (Non-reasoning   49.8  (no Nous route)                           
GPT-5.1 (High)                 49.4  openai/gpt-5.1:batch                          0.62    5.00     28.7
Nemotron 3 Ultra 550B A55B (   49.3  (no Nous route)                           
Gemini 3.5 Flash-Lite          49.3  google/gemini-3.5-flash-lite:batch            0.15    1.25    116.0
Muse Glimmer (High)            49.0  meta/muse-glimmer-30b                         0.30    1.10     98.0
Qwen3.5 397B A17B (Reasoning   48.2  qwen/qwen3.5-397b-a17b                        0.39    2.34     54.9
Mistral Medium 3.5             46.9  mistralai/mistral-medium-3-5:batch            0.75    3.75     31.3
Kimi K2.5 (Reasoning)          46.8  moonshotai/kimi-k2.5                          0.45    2.25     52.0
Gemini 2.5 Pro Preview (Mar'   46.7  google/gemini-2.5-pro:batch                   0.62    5.00     27.2
Qwen3.6 27B (Non-reasoning)    46.6  qwen/qwen3.6-27b                              0.30    2.00     64.3
GLM-5.2 (Non-reasoning)        46.5  z-ai/glm-5                                    0.60    1.92     50.0
GLM-4.6 (Reasoning)            45.8  z-ai/glm-4.6                                  0.43    1.75     60.3
Qwen3.5 122B A10B (Reasoning   45.7  qwen/qwen3.5-122b-a10b                        0.26    2.08     63.9
GLM-4.7 (Reasoning)            45.3  z-ai/glm-4.7                                  0.40    1.75     61.4
LongCat 2.0                    45.3  meituan/longcat-2.0:free                      0.00    0.00  45300.0
Solar Open2 250B               45.0  (no Nous route)                           
Qwen3.8 27B (Non-reasoning)    44.6  qwen/qwen3.8-27b                              0.02    4.35     40.3
DeepSeek V3.2 (Reasoning)      44.2  deepseek/deepseek-v3.2                        0.21    0.31    188.9
GPT-5.6 Luna (Low)             44.2  openai/gpt-5.6-luna:batch                     0.10    0.60    196.4
Claude 4.5 Haiku (Reasoning)   43.9  (no Nous route)                           
DeepSeek V3.1 Terminus (Reas   43.5  deepseek/deepseek-v3.1-terminus               0.27    1.00     96.1
Gemma 4 31B (Reasoning)        43.4  google/gemma-4-31b-it                         0.09    0.34    284.6
Qwen3.5 122B A10B (Non-reaso   43.3  qwen/qwen3.5-122b-a10b                        0.26    2.08     60.6
Ring-2.6-1T                    42.8  (no Nous route)                           
Grok 4.3 (High)                42.2  x-ai/grok-4.3:batch                           1.00    2.00     33.8
Qwen3.6 35B A3B (Reasoning)    41.9  qwen/qwen3.6-35b-a3b                          0.05    0.70    197.2
K-EXAONE 2.0 0803              40.6  (no Nous route)                           
o1                             39.7  openai/o1                                    15.00   60.00      1.5
Step 3.7 Flash                 39.6  stepfun/step-3.7-flash:free                   0.00    0.00  39600.0
GPT-5.5 Instant (June 2026)    39.4  openai/gpt-5.5:batch                          2.50   15.00      7.0
Gemma 4 26B A4B (Reasoning)    39.3  google/gemma-4-26b-a4b-it                     0.04    0.22    454.3
GPT-5.6 Luna (Non-reasoning)   39.3  openai/gpt-5.6-luna:batch                     0.10    0.60    174.7
A.X-K2                         38.8  (no Nous route)                           
K2 Horizon 7B                  38.6  (no Nous route)                           
GPT-5 (High)                   37.8  openai/gpt-5:batch                            0.62    5.00     22.0
Nemotron 3 Super 120B A12B (   37.7  (no Nous route)                           
Claude 4 Sonnet (Reasoning)    37.6  (no Nous route)                           
Qwen3.5 35B A3B (Non-reasoni   37.0  qwen/qwen3.5-35b-a3b                          0.08    0.75    149.5
North Mini Code                36.5  (no Nous route)                           
Claude 3.7 Sonnet (Reasoning   36.4  (no Nous route)                           
Qwen3 Coder Next               36.2  qwen/qwen3-coder-next                         0.12    0.80    124.8
Grok 4.3 (Non-reasoning)       35.2  x-ai/grok-4.3:batch                           1.00    2.00     28.2
Gemini 3.1 Flash-Lite          34.7  google/gemini-3.1-flash-lite:batch            0.12    0.75    123.4
Nova 2.0 Pro Preview (Medium   34.0  (no Nous route)                           
o1-preview                     34.0  openai/o1                                    15.00   60.00      1.3
Gemini 2.5 Pro                 33.3  google/gemini-2.5-pro:batch                   0.62    5.00     19.4
Gemma 4 31B (Non-reasoning)    33.2  (no Nous route)                           
G9v3-39A5B                     33.1  (no Nous route)                           
K-EXAONE (Reasoning)           32.1  (no Nous route)                           
Devstral 2                     31.3  mistralai/devstral-2512                       0.40    2.00     39.1
Mercury 2                      31.1  inception/mercury-2                           0.25    0.75     82.9
Gemma 4 12B (Reasoning)        31.0  (no Nous route)                           
gpt-oss-120b (High)            30.4  openai/gpt-oss-120b:batch                     0.03    0.14    540.9
Claude 3.5 Sonnet (Oct '24)    30.2  (no Nous route)                           
Granite 4.2 30B                29.9  (no Nous route)                           
Devstral Small 2               29.3  (no Nous route)                           
Qwen3.5 9B (Reasoning)         28.7  qwen/qwen3.5-9b                               0.08    0.13    310.3
Qwen3.6 35B A3B (Non-reasoni   28.1  qwen/qwen3.6-35b-a3b                          0.05    0.70    132.2
Command A+                     27.8  cohere/command-a-plus                         0.30    1.50     46.3
Nemotron 3.5 Lightning         26.8  nvidia/nemotron-3.5-lightning                 0.04    0.18    360.9
Mistral Small 4 (Reasoning)    26.6  (no Nous route)                           
Ling 3.0 Tiny                  26.5  (no Nous route)                           
Mistral Small 3.1              26.3  mistralai/mistral-small-3.1-24b-instruct      0.35    0.55     65.4
K2 Horizon 3.7B                26.1  (no Nous route)                           
Claude 3.5 Sonnet (June '24)   26.0  (no Nous route)                           
Nova 2.0 Pro Preview (Low)     25.9  (no Nous route)                           
Trinity Large Thinking         25.8  arcee-ai/trinity-large-thinking               0.25    0.80     66.6
Gemini 2.0 Pro Experimental    25.5  (no Nous route)                           
Nemotron Cascade 2 30B A3B     25.3  (no Nous route)                           
Ling 2.6 Flash                 25.3  (no Nous route)                           
DeepSeek R1 (Jan '25)          24.6  deepseek/deepseek-r1                          0.70    2.50     21.4
GPT-4o (May '24)               24.2  openai/gpt-4o-2024-05-13                      5.00   15.00      3.2
Gemini 2.0 Flash Thinking Ex   24.1  (no Nous route)                           
EXAONE 4.5 33B (Reasoning)     23.6  (no Nous route)                           
Gemini 1.5 Pro (Sep '24)       23.6  (no Nous route)                           
Qwen3.5 9B (Non-reasoning)     23.5  qwen/qwen3.5-9b                               0.08    0.13    254.1
HyperNova 60B 2605 (High, Ba   23.2  (no Nous route)                           
Nova 2.0 Lite (High)           23.0  (no Nous route)                           
DeepSeek V3 (Dec '24)          23.0  deepseek/deepseek-v3.2                        0.21    0.31     98.3
Qwen3.5 4B (Reasoning)         22.6  (no Nous route)                           
Granite 4.2 8B                 22.4  ibm-granite/granite-4.2-8b                    0.06    0.25    208.4
Qwen3 235B A22B 2507 (Reason   22.1  qwen/qwen3-235b-a22b                          0.45    1.82     27.8
GPT-4 Turbo                    21.5  openai/gpt-4-turbo:batch                      5.00   15.00      2.9
Magistral Medium 1.2           21.3  (no Nous route)                           
gpt-oss-120b (Low)             21.2  openai/gpt-oss-120b:batch                     0.03    0.14    377.2
DeepSeek V3 0324               21.2  (no Nous route)                           
K2 Think V2                    21.0  (no Nous route)                           
Nova 2.0 Pro Preview (Non-re   20.9  (no Nous route)                           
gpt-oss-20b (High)             20.7  openai/gpt-oss-20b                            0.02    0.09    575.0
Mistral Medium 3.1             20.5  mistralai/mistral-medium-3.1:batch            0.20    1.00     51.2
Qwen3.5 4B (Non-reasoning)     20.3  (no Nous route)                           
GPT-4.1 mini                   20.2  openai/gpt-4.1-mini:batch                     0.20    0.80     57.7
Mistral Large 3                20.1  mistralai/mistral-large                       2.00    6.00      6.7
Gemini 1.5 Pro (May '24)       19.8  (no Nous route)                           
DiffusionGemma 26B A4B         19.7  (no Nous route)                           
Claude 3 Opus                  19.5  (no Nous route)                           
Gemini 1.0 Ultra               17.6  (no Nous route)                           
Granite 4.2 3B                 17.5  (no Nous route)                           
Qwen3 Next 80B A3B (Reasonin   17.4  (no Nous route)                           
o3-mini (High)                 16.3  openai/o3-mini-high                           1.10    4.40      8.5
Llama 4 Maverick               16.3  meta-llama/llama-4-maverick                   0.19    0.65     53.7
Solar Pro 3                    16.2  upstage/solar-pro-3                           0.15    0.60     61.7
Claude 3.5 Haiku               15.9  (no Nous route)                           
GPT-5 mini (High)              15.6  openai/gpt-5-mini:batch                       0.12    1.00     45.4
Qwen3 32B (Reasoning)          15.3  qwen/qwen3-32b                                0.08    0.28    117.7
Magistral Small 1.2            14.7  (no Nous route)                           
MiniCPM5-2B                    14.5  (no Nous route)                           
NVIDIA Nemotron 3 Nano 30B A   14.4  (no Nous route)                           
Celeris-1                      14.4  (no Nous route)                           
Ministral 3 14B                14.4  (no Nous route)                           
Claude 2.1                     14.0  (no Nous route)                           
Nemotron 3 Nano Omni 30B A3B   13.8  (no Nous route)                           
Qwen3 14B (Reasoning)          13.8  qwen/qwen3-14b                                0.10    0.22    106.2
GPT-4                          13.1  openai/gpt-4                                 30.00   60.00      0.3
Claude 2.0                     12.9  (no Nous route)                           
Mistral Small 3.2              12.5  mistralai/mistral-small-3.2-24b-instruct      0.07    0.20    117.6
Qwen3 30B A3B 2507 (Reasonin   12.1  qwen/qwen3-30b-a3b                            0.12    0.50     56.3
Llama 3.3 Instruct 70B         11.9  (no Nous route)                           
GPT-4o mini                    11.4  openai/gpt-4o-mini:batch                      0.07    0.30     86.9
GPT-4.1 nano                   11.1  openai/gpt-4.1-nano:batch                     0.05    0.20    126.9
GPT-3.5 Turbo                  10.7  openai/gpt-3.5-turbo:batch                    0.25    0.75     28.5
Granite 4.1 30B                10.4  (no Nous route)                           
Gemma 3 27B Instruct           10.1  google/gemma-3-27b-it                         0.08    0.16    101.0
G9v3-3B                         9.9  (no Nous route)                           
Ministral 3 8B                  9.7  (no Nous route)                           
Nanbeige4.1-3B                  9.6  (no Nous route)                           
Granite 4.1 8B                  9.5  (no Nous route)                           
Qwen3 8B (Reasoning)            9.0  qwen/qwen3-8b                                 0.12    0.45     44.7
Llama 4 Scout                   8.2  meta-llama/llama-4-scout                      0.10    0.30     54.7
NVIDIA Nemotron 3 Nano 4B       8.0  (no Nous route)                           
Claude Instant                  7.8  (no Nous route)                           
LFM2.5-2.6B                     7.7  (no Nous route)                           
Gemma 4 E2B (Reasoning)         7.2  (no Nous route)                           
Gemma 3 12B Instruct            5.8  google/gemma-3-12b-it                         0.05    0.15     77.3
Llama 3.1 Instruct 8B           5.4  (no Nous route)                           
Ministral 3 3B                  4.8  (no Nous route)                           
Granite 4.1 3B                  4.7  (no Nous route)                           
PALM-2                          4.6  (no Nous route)                           
Phi-4 Mini Instruct             3.8  (no Nous route)                           
K2 Horizon 0.9B                 3.4  (no Nous route)                           
Gemma 3n E4B Instruct           3.2  (no Nous route)                           
Qwen3.5 2B (Reasoning)          2.9  (no Nous route)                           
Gemma 3 4B Instruct             2.7  (no Nous route)                           
Qwen3.5 2B (Non-reasoning)      2.4  (no Nous route)                           
Qwen3.5 0.8B (Non-reasoning)    1.2  (no Nous route)                           
MiniCPM-V 4.6 1.3B              0.7  (no Nous route)                           
Qwen3.5 0.8B (Reasoning)        0.0  (no Nous route)                           
