# Terminal-Bench 2.1 (Vals AI) → Terminal-Bench Hard

Calibration · https://benchgap.net/calibration/1573 · data retrieved 2026-10-29

Terminal-Bench Hard is estimated from Terminal-Bench 2.1 (Vals AI) with a Hill curve fitted on 14 models measured on both: y = 0.4319 + (0.9323 − 0.4319)·x^6.00 / (0.87197^6.00 + x^6.00), R² = 0.81, cross-validated error 3.8 pp. It is used for 46 estimates.

| Estimated model | Terminal-Bench 2.1 (Vals AI) | Terminal-Bench Hard | Source |
|---|---|---|---|
| [GPT-6 Astra](https://benchgap.net/model/gpt-6-astra.md) | 87.3% | 68.3% | estimated ± 3.8 pp, medium confidence |
| [Claude Fable 5.1](https://benchgap.net/model/claude-fable-5-1.md) | 85.0% | 66.3% | estimated ± 3.8 pp, high confidence |
| [Claude Opus 5](https://benchgap.net/model/claude-opus-5.md) | 84.6% | 65.9% | estimated ± 3.8 pp, high confidence |
| [Gemini 3.8 Flash](https://benchgap.net/model/gemini-3-8-flash.md) | 81.3% | 63.0% | estimated ± 3.8 pp, high confidence |
| [Kimi K3](https://benchgap.net/model/kimi-k3.md) | 80.9% | 62.7% | estimated ± 3.8 pp, high confidence |
| [GPT-5.6 Luna](https://benchgap.net/model/gpt-5-6-luna.md) | 79.0% | 61.0% | estimated ± 3.8 pp, high confidence |
| [Grok 4.6](https://benchgap.net/model/grok-4-6.md) | 78.3% | 60.4% | estimated ± 3.8 pp, high confidence |
| [Gemini 3.7 Flash](https://benchgap.net/model/gemini-3-7-flash.md) | 77.5% | 59.7% | estimated ± 3.8 pp, high confidence |
| [GPT-5.6 Terra](https://benchgap.net/model/gpt-5-6-terra.md) | 77.5% | 59.7% | estimated ± 3.8 pp, high confidence |
| [Claude Sonnet 5](https://benchgap.net/model/claude-sonnet-5.md) | 74.5% | 57.2% | estimated ± 3.8 pp, high confidence |
| [DeepSeek V4.1 Flash 1017](https://benchgap.net/model/deepseek-v4-1-flash.md) | 74.5% | 57.2% | estimated ± 3.8 pp, high confidence |
| [Gemini 3.5 Flash](https://benchgap.net/model/gemini-3-5-flash.md) | 74.2% | 57.0% | estimated ± 3.8 pp, high confidence |
| [Gemini 3.6 Flash](https://benchgap.net/model/gemini-3-6-flash.md) | 73.8% | 56.6% | estimated ± 3.8 pp, high confidence |
| [Grok 4.7](https://benchgap.net/model/grok-4-7.md) | 73.4% | 56.3% | estimated ± 3.8 pp, high confidence |
| [Muse Spark 1.3](https://benchgap.net/model/muse-spark-1-3.md) | 72.3% | 55.5% | estimated ± 3.8 pp, high confidence |
| [GLM-5.3](https://benchgap.net/model/glm-5-3.md) | 71.5% | 54.9% | estimated ± 3.8 pp, high confidence |
| [Muse Spark 1.2](https://benchgap.net/model/muse-spark-1-2.md) | 69.7% | 53.5% | estimated ± 3.8 pp, high confidence |
| [Muse Spark 1.1](https://benchgap.net/model/muse-spark-1-1.md) | 69.3% | 53.3% | estimated ± 3.8 pp, high confidence |
| [Grok 4.5](https://benchgap.net/model/grok-4-5.md) | 67.8% | 52.2% | estimated ± 3.8 pp, high confidence |
| [Qwen3.8 Max](https://benchgap.net/model/qwen3-8-max.md) | 67.4% | 52.0% | estimated ± 3.8 pp, high confidence |
| [DeepSeek V4 Flash 0731](https://benchgap.net/model/deepseek-v4-flash-0731.md) | 67.0% | 51.7% | estimated ± 3.8 pp, high confidence |
| [GLM-5.3-Flash](https://benchgap.net/model/glm-5-3-flash.md) | 62.9% | 49.4% | estimated ± 3.8 pp, high confidence |
| [MiMo-V2.5](https://benchgap.net/model/mimo-v2-5.md) | 60.7% | 48.3% | estimated ± 3.8 pp, high confidence |
| [Composer 2.5](https://benchgap.net/model/composer-2-5.md) | 58.4% | 47.3% | estimated ± 3.8 pp, high confidence |
| [Qwen3.8-27B](https://benchgap.net/model/qwen3-8-27b.md) | 58.4% | 47.3% | estimated ± 3.8 pp, high confidence |
| [Claude Sonnet 4.6](https://benchgap.net/model/claude-sonnet-4-6.md) | 57.3% | 46.9% | estimated ± 3.8 pp, high confidence |
| [GLM-5.1](https://benchgap.net/model/glm-5-1.md) | 56.9% | 46.8% | estimated ± 3.8 pp, high confidence |
| [Hy4 preview](https://benchgap.net/model/hy4-preview.md) | 55.1% | 46.2% | estimated ± 3.8 pp, high confidence |
| [Inkling-Small](https://benchgap.net/model/inkling-small.md) | 55.1% | 46.2% | estimated ± 3.8 pp, high confidence |
| [DeepSeek V4 Pro 0813](https://benchgap.net/model/deepseek-v4-pro-0813.md) | 54.7% | 46.1% | estimated ± 3.8 pp, high confidence |
| [Gemini 3 Flash](https://benchgap.net/model/gemini-3-flash.md) | 53.9% | 45.8% | estimated ± 3.8 pp, high confidence |
| [MiniMax M3](https://benchgap.net/model/minimax-m3.md) | 53.6% | 45.7% | estimated ± 3.8 pp, high confidence |
| [Nemotron 3 Ultra](https://benchgap.net/model/nemotron-3-ultra.md) | 50.9% | 45.1% | estimated ± 3.8 pp, medium confidence |
| [Gemini 3.5 Flash-Lite](https://benchgap.net/model/gemini-3-5-flash-lite.md) | 50.2% | 44.9% | estimated ± 3.8 pp, medium confidence |
| [Ling 3.0 Flash](https://benchgap.net/model/ling-3-0-flash.md) | 50.2% | 44.9% | estimated ± 3.8 pp, medium confidence |
| [MiniMax M2.7](https://benchgap.net/model/minimax-m2-7.md) | 48.7% | 44.7% | estimated ± 3.8 pp, medium confidence |
| [Inkling](https://benchgap.net/model/inkling.md) | 47.6% | 44.5% | estimated ± 3.8 pp, medium confidence |
| [Grok 4.20](https://benchgap.net/model/grok-4-20-beta.md) | 44.2% | 44.0% | estimated ± 3.8 pp, medium confidence |
| [Claude Haiku 4.5](https://benchgap.net/model/claude-haiku-4-5.md) | 43.8% | 44.0% | estimated ± 3.8 pp, medium confidence |
| [Grok 4.3](https://benchgap.net/model/grok-4-3.md) | 41.9% | 43.8% | estimated ± 3.8 pp, medium confidence |
| [GPT-5.4 nano](https://benchgap.net/model/gpt-5-4-nano.md) | 41.6% | 43.8% | estimated ± 3.8 pp, medium confidence |
| [Mistral Medium 3.5 128B](https://benchgap.net/model/mistral-medium-3-5-128b.md) | 39.0% | 43.6% | estimated ± 3.8 pp, medium confidence |
| [Mercury 2.5](https://benchgap.net/model/mercury-2-5.md) | 34.5% | 43.4% | estimated ± 3.8 pp, medium confidence |
| [Gemini 3.1 Flash-Lite](https://benchgap.net/model/gemini-3-1-flash-lite.md) | 34.1% | 43.4% | estimated ± 3.8 pp, medium confidence |
| [Laguna M.1](https://benchgap.net/model/laguna-m-1.md) | 34.1% | 43.4% | estimated ± 3.8 pp, medium confidence |
| [Laguna XS.2](https://benchgap.net/model/laguna-xs-2.md) | 25.8% | 43.2% | estimated ± 3.8 pp, medium confidence |

## Fitted on

| Model | Terminal-Bench 2.1 (Vals AI) | Terminal-Bench Hard |
|---|---|---|
| [Claude Fable 5](https://benchgap.net/model/claude-fable.md) | 80.5% | 62.9% |
| [Claude Opus 4.7](https://benchgap.net/model/claude-opus-4-7.md) | 68.5% | 54.5% |
| [Claude Opus 4.8](https://benchgap.net/model/claude-opus-4-8.md) | 71.9% | 58.3% |
| [Gemini 3.1 Pro](https://benchgap.net/model/gemini-3-1-pro.md) | 70.8% | 53.8% |
| [GLM-5.2](https://benchgap.net/model/glm-5-2.md) | 67.8% | 50.8% |
| [GPT-5.4 mini](https://benchgap.net/model/gpt-5-4-mini.md) | 54.7% | 52.3% |
| [GPT-5.5](https://benchgap.net/model/gpt-5-5.md) | 76.4% | 60.6% |
| [GPT-5.6 Sol](https://benchgap.net/model/gpt-5-6-sol.md) | 85.8% | 65.9% |
| [Kimi K2.6](https://benchgap.net/model/kimi-2-6.md) | 53.6% | 43.9% |
| [Kimi K2.7 Code](https://benchgap.net/model/kimi-k2-7-code.md) | 67.0% | 44.7% |
| [MiMo-V2.5-Pro](https://benchgap.net/model/mimo-v2-5-pro.md) | 57.3% | 43.2% |
| [Qwen3.6 Plus](https://benchgap.net/model/qwen3-6-plus.md) | 53.2% | 43.9% |
| [Qwen3.7 Max](https://benchgap.net/model/qwen3-7-max.md) | 61.0% | 50.8% |
| [Qwen3.7 Plus](https://benchgap.net/model/qwen3-7-plus.md) | 52.8% | 47.0% |