MODEL DIRECTORY

Follow the model.
See the evidence.

Benchmark results, evaluation settings and disclosed training methods—collected in one place for each model.

3,335 reported models and checkpoint variants. Scores keep their original units and sources.

510 models · page 1 of 11

DeepSeek

DeepSeek V4.1 Flash 1017

74 benchmarks · 171 observations

Explore results →

NVIDIA

Nemotron 3.5 Super VL

19 benchmarks · 23 observations

Explore results →

StepFun

Step 5 Preview

44 benchmarks · 52 observations

Explore results →

StepFun

Step 5 Preview (unknown)

3 benchmarks · 3 observations

Explore results →

DeepSeek

DeepSeek V4.1 Flash

60 benchmarks · 71 observations

Explore results →

DeepSeek

DeepSeek V4.1 Flash (none)

4 benchmarks · 4 observations

Explore results →

DeepSeek

DeepSeek V4.1 Flash (unknown)

4 benchmarks · 4 observations

Explore results →

Qwen

Qwen3.8 Max (0902)

47 benchmarks · 55 observations

Explore results →

Qwen

Qwen3.8 Max (0902) (unknown)

2 benchmarks · 2 observations

Explore results →

Qwen

Qwen3.8-Flash-Next

3 benchmarks · 3 observations

Explore results →

Qwen

Qwen3.8 Flash

2 benchmarks · 2 observations

Explore results →

DeepSeek

DeepSeek V4 Flash Vision

2 benchmarks · 2 observations

Explore results →

Qwen

Qwen3.8 27B

54 benchmarks · 77 observations

Explore results →

DeepSeek

DeepSeek V4 Pro 0813

46 benchmarks · 55 observations

Explore results →

DeepSeek

DeepSeek V4 Pro 0813 (none)

4 benchmarks · 4 observations

Explore results →

DeepSeek

DeepSeek V4 Pro 0813 (unknown)

4 benchmarks · 4 observations

Explore results →

Qwen

Qwen3.8 2.4T A95B

5 benchmarks · 5 observations

Explore results →

Qwen

Qwen3.8 Max

48 benchmarks · 112 observations

Explore results →

Qwen

Qwen3.8 Max (unknown)

5 benchmarks · 6 observations

Explore results →

DeepSeek

DeepSeek V4 Flash 0731

41 benchmarks · 47 observations

Explore results →

DeepSeek

DeepSeek V4 Flash 0731 (unknown)

3 benchmarks · 3 observations

Explore results →

DeepSeek

DeepSeek V4 Flash 0731 (none)

2 benchmarks · 2 observations

Explore results →

Qwen

Qwen3.7 Flash

5 benchmarks · 9 observations

Explore results →

NVIDIA

Nemotron 3 Ultra

27 benchmarks · 28 observations

Explore results →

NVIDIA

Nemotron 3 Ultra 550B A55B (Reasoning)

3 benchmarks · 3 observations

Explore results →

Qwen

Qwen3.7 Plus

26 benchmarks · 30 observations

Explore results →

StepFun

Step 3.7 Flash

4 benchmarks · 4 observations

Explore results →

Qwen

Qwen3.7 Max

40 benchmarks · 43 observations

Explore results →

Qwen

Qwen3.6 Flash (2026-04-16)

11 benchmarks · 13 observations

Explore results →

Qwen

Qwen3.6 Flash

2 benchmarks · 2 observations

Explore results →

DeepSeek

DeepSeek V4 Pro

28 benchmarks · 42 observations

Explore results →

DeepSeek

DeepSeek V4 Flash

11 benchmarks · 16 observations

Explore results →

DeepSeek

DeepSeek V4 Pro 0424

5 benchmarks · 5 observations

Explore results →

DeepSeek

DeepSeek V4 Flash 0420

2 benchmarks · 2 observations

Explore results →

DeepSeek

DeepSeek V4 (unknown)

1 benchmarks · 1 observations

Explore results →

Qwen

Qwen3.6 27B

15 benchmarks · 19 observations

Explore results →

Qwen

Qwen3.6 Max Preview

17 benchmarks · 18 observations

Explore results →

Qwen

Qwen3.6 35B A3B (Reasoning)

1 benchmarks · 1 observations

Explore results →

Qwen

Qwen3.6 35B-A3B

13 benchmarks · 18 observations

Explore results →

Qwen

Qwen3.6 Plus

27 benchmarks · 28 observations

Explore results →

Qwen

Qwen3.6 Plus (2026-04-02)

15 benchmarks · 16 observations

Explore results →

NVIDIA

Nemotron Cascade 2 30B A3B

1 benchmarks · 1 observations

Explore results →

NVIDIA

Nemotron 3 Super

4 benchmarks · 4 observations

Explore results →

Qwen

Qwen3.5 Flash

23 benchmarks · 27 observations

Explore results →

Qwen

Qwen3.5 Flash (hosted 35B-A3B)

1 benchmarks · 1 observations

Explore results →

Qwen

Qwen3.5-35B-A3B

9 benchmarks · 11 observations

Explore results →

Qwen

Qwen3.5 9B

8 benchmarks · 10 observations

Explore results →

Qwen

Qwen3.5 27B

6 benchmarks · 6 observations

Explore results →