Calibration
AA Coding Index → React Native Evals
React Native Evals is estimated from AA Coding Index with a linear curve fitted on 8 models measured on both: y = 0.2513·x + 0.6407, R² = 0.74, cross-validated error 3.5 pp. It is used for 50 estimates.
| Estimated model | AA Coding Index | React Native Evals | Source |
|---|---|---|---|
| Apodex 1.1 | 60.8% | 79.3% | estimated ± 3.5 pp, high confidence |
| Apodex 1.1 Mini | 60.8% | 79.3% | estimated ± 3.5 pp, high confidence |
| Celeris-1 | 14.4% | 67.7% | estimated ± 3.5 pp, medium confidence |
| Claude 3 Opus | 19.5% | 69.0% | estimated ± 3.5 pp, medium confidence |
| Claude Fable 5.1 | 81.6% | 84.6% | estimated ± 3.5 pp, medium confidence |
| Claude Opus 4.7 (Adaptive) | 73.6% | 82.6% | estimated ± 3.5 pp, high confidence |
| Command A+ | 27.9% | 71.1% | estimated ± 3.5 pp, high confidence |
| DeepSeek V3 | 23.0% | 69.9% | estimated ± 3.5 pp, high confidence |
| Gemini 1.5 Pro | 23.6% | 70.0% | estimated ± 3.5 pp, high confidence |
| Gemma 3 27B | 10.1% | 66.6% | estimated ± 3.5 pp, medium confidence |
| Gemma 4 12B | 31.0% | 71.9% | estimated ± 3.5 pp, high confidence |
| Gemma 4 26B A4B | 39.3% | 74.0% | estimated ± 3.5 pp, high confidence |
| Gemma 4 E2B | 7.2% | 65.9% | estimated ± 3.5 pp, medium confidence |
| Gemma 4 E4B | 9.4% | 66.4% | estimated ± 3.5 pp, medium confidence |
| GPT-4.1 mini | 20.2% | 69.2% | estimated ± 3.5 pp, medium confidence |
| GPT-4.1 nano | 11.1% | 66.9% | estimated ± 3.5 pp, medium confidence |
| GPT-4 Turbo | 21.5% | 69.5% | estimated ± 3.5 pp, high confidence |
| GPT-4o mini | 11.4% | 66.9% | estimated ± 3.5 pp, medium confidence |
| GPT-6 Astra | 76.9% | 83.4% | estimated ± 3.5 pp, medium confidence |
| Granite 4.2 8B | 22.4% | 69.7% | estimated ± 3.5 pp, high confidence |
| Hy3 | 58.8% | 78.8% | estimated ± 3.5 pp, high confidence |
| Hy3 Preview | 58.8% | 78.8% | estimated ± 3.5 pp, high confidence |
| K-Exaone | 32.1% | 72.1% | estimated ± 3.5 pp, high confidence |
| LFM2.5-2.6B | 7.7% | 66.0% | estimated ± 3.5 pp, medium confidence |
| Ling 2.6 Flash | 25.3% | 70.4% | estimated ± 3.5 pp, high confidence |
| Ling 3.0 Flash FP8 | 50.6% | 76.8% | estimated ± 3.5 pp, high confidence |
| Llama 4 Maverick | 16.3% | 68.2% | estimated ± 3.5 pp, medium confidence |
| Llama 4 Scout | 8.2% | 66.1% | estimated ± 3.5 pp, medium confidence |
| MiMo-V2-Flash | 49.8% | 76.6% | estimated ± 3.5 pp, high confidence |
| Mistral Large 3 | 20.1% | 69.1% | estimated ± 3.5 pp, medium confidence |
| Mistral Small 4 | 26.6% | 70.8% | estimated ± 3.5 pp, high confidence |
| Mistral Small 4 (Reasoning) | 26.6% | 70.8% | estimated ± 3.5 pp, high confidence |
| Muse Glimmer 30B | 49.0% | 76.4% | estimated ± 3.5 pp, high confidence |
| Muse Spark 1.3 | 75.8% | 83.1% | estimated ± 3.5 pp, medium confidence |
| Nemotron 3.5 Lightning 30B A3B NVFP4 | 26.8% | 70.8% | estimated ± 3.5 pp, high confidence |
| Nemotron 3 Nano 30B | 14.4% | 67.7% | estimated ± 3.5 pp, medium confidence |
| Nemotron 3 Nano Omni 30B A3B | 13.8% | 67.5% | estimated ± 3.5 pp, medium confidence |
| Nemotron 3 Super 100B | 37.7% | 73.6% | estimated ± 3.5 pp, high confidence |
| o1 | 39.7% | 74.1% | estimated ± 3.5 pp, high confidence |
| o1-preview | 34.1% | 72.6% | estimated ± 3.5 pp, high confidence |
| Quasar 438B | 61.2% | 79.5% | estimated ± 3.5 pp, high confidence |
| Qwen3.5-122B-A10B | 45.7% | 75.6% | estimated ± 3.5 pp, high confidence |
| Qwen3.6-35B-A3B | 41.9% | 74.6% | estimated ± 3.5 pp, high confidence |
| Qwen3.7 Plus | 55.9% | 78.1% | estimated ± 3.5 pp, high confidence |
| Qwen3.8-Flash-Next | 73.1% | 82.4% | estimated ± 3.5 pp, high confidence |
| Qwen3.8 Max Preview | 71.8% | 82.1% | estimated ± 3.5 pp, high confidence |
| Step 3.7 Flash | 39.6% | 74.0% | estimated ± 3.5 pp, high confidence |
| Trinity-Large-Preview | 25.8% | 70.5% | estimated ± 3.5 pp, high confidence |
| Trinity-Large-Thinking | 25.8% | 70.5% | estimated ± 3.5 pp, high confidence |
| Ultravox v0.6 Llama 3.3 70B | 11.9% | 67.1% | estimated ± 3.5 pp, medium confidence |