# FrontierMath v2 (Tiers 1-3) → FrontierMath v2 (Tier 4)

Calibration · https://benchgap.net/calibration/frontiermathv2tiers13/current/to/frontiermathv2tier4/current · data retrieved 2026-11-18

FrontierMath v2 (Tier 4) is estimated from FrontierMath v2 (Tiers 1-3) with a inverse Michaelis–Menten curve fitted on 56 models measured on both: y = 0.34536·(x − 0.2178) / (0.2178 + 0.9434 − x), R² = 0.99, cross-validated error 4.2 pp. It is used for 8 estimates.

| Estimated model | FrontierMath v2 (Tiers 1-3) | FrontierMath v2 (Tier 4) | Source |
|---|---|---|---|
| [o3](https://benchgap.net/model/o3.md) | 33.3% | 4.8% | estimated ± 4.2 pp, high confidence |
| [Gemini 3.1 Flash-Lite](https://benchgap.net/model/gemini-3-1-flash-lite.md) | 27.7% | 2.3% | estimated ± 4.2 pp, high confidence |
| [GPT-4.1](https://benchgap.net/model/gpt-4-1.md) | 6.0% | 0.0% | estimated ± 4.2 pp, medium confidence |
| [GPT-4.1 mini](https://benchgap.net/model/gpt-4-1-mini.md) | 6.7% | 0.0% | estimated ± 4.2 pp, medium confidence |
| [GPT-4o](https://benchgap.net/model/gpt-4o.md) | 0.4% | 0.0% | estimated ± 4.2 pp, medium confidence |
| [o1](https://benchgap.net/model/o1.md) | 14.7% | 0.0% | estimated ± 4.2 pp, medium confidence |
| [Qwen3.5 Flash](https://benchgap.net/model/qwen3-5-flash.md) | 18.2% | 0.0% | estimated ± 4.2 pp, medium confidence |
| [Qwen3.7 Flash](https://benchgap.net/model/qwen3-7-flash.md) | 19.3% | 0.0% | estimated ± 4.2 pp, medium confidence |

## Fitted on

| Model | FrontierMath v2 (Tiers 1-3) | FrontierMath v2 (Tier 4) |
|---|---|---|
| [Claude Fable 5](https://benchgap.net/model/claude-fable.md) | 87.0% | 90.2% |
| [Claude Fable 5.1](https://benchgap.net/model/claude-fable-5-1.md) | 90.2% | 87.8% |
| [Claude Fable 5.2](https://benchgap.net/model/claude-fable-5-2.md) | 95.1% | 100.0% |
| [Claude Haiku 5.5](https://benchgap.net/model/claude-haiku-5-5.md) | 75.1% | 46.3% |
| [Claude Opus 4.5](https://benchgap.net/model/claude-opus-4-5.md) | 34.4% | 4.9% |
| [Claude Opus 4.6](https://benchgap.net/model/claude-opus-4-6.md) | 66.0% | 26.8% |
| [Claude Opus 4.7](https://benchgap.net/model/claude-opus-4-7.md) | 70.2% | 31.7% |
| [Claude Opus 4.8](https://benchgap.net/model/claude-opus-4-8.md) | 80.0% | 56.1% |
| [Claude Opus 5](https://benchgap.net/model/claude-opus-5.md) | 85.6% | 73.2% |
| [Claude Opus 5.5](https://benchgap.net/model/claude-opus-5-5.md) | 91.2% | 95.0% |
| [Claude Sonnet 4.5](https://benchgap.net/model/claude-sonnet-4-5.md) | 23.9% | 2.4% |
| [Claude Sonnet 5](https://benchgap.net/model/claude-sonnet-5.md) | 65.6% | 29.3% |
| [Claude Sonnet 5.5](https://benchgap.net/model/claude-sonnet-5-5.md) | 88.8% | 80.5% |
| [DeepSeek V4.1 Flash 1017](https://benchgap.net/model/deepseek-v4-1-flash.md) | 67.4% | 26.8% |
| [DeepSeek V4 Pro](https://benchgap.net/model/deepseek-v4-pro.md) | 64.6% | 26.8% |
| [Gemini 2.5 Pro](https://benchgap.net/model/gemini-2-5-pro.md) | 24.6% | 0.0% |
| [Gemini 3.1 Pro](https://benchgap.net/model/gemini-3-1-pro.md) | 59.6% | 26.8% |
| [Gemini 3.5 Flash](https://benchgap.net/model/gemini-3-5-flash.md) | 62.8% | 26.8% |
| [Gemini 3.5 Flash-Lite](https://benchgap.net/model/gemini-3-5-flash-lite.md) | 26.0% | 0.0% |
| [Gemini 3.6 Flash](https://benchgap.net/model/gemini-3-6-flash.md) | 58.9% | 22.0% |
| [Gemini 3.7 Flash](https://benchgap.net/model/gemini-3-7-flash.md) | 71.6% | 36.6% |
| [Gemini 3.8 Flash](https://benchgap.net/model/gemini-3-8-flash.md) | 68.4% | 22.0% |
| [Gemini 3 Flash](https://benchgap.net/model/gemini-3-flash.md) | 51.2% | 17.1% |
| [Gemini 4 Argon](https://benchgap.net/model/gemini-4-argon.md) | 86.0% | 78.0% |
| [Gemini 4 Carbon](https://benchgap.net/model/gemini-4-carbon.md) | 90.5% | 95.1% |
| [GLM-5.2](https://benchgap.net/model/glm-5-2.md) | 59.2% | 29.3% |
| [GLM-5.3](https://benchgap.net/model/glm-5-3.md) | 68.8% | 29.3% |
| [GLM-5.3-Flash](https://benchgap.net/model/glm-5-3-flash.md) | 55.8% | 17.1% |
| [GPT-5.2](https://benchgap.net/model/gpt-5-2.md) | 67.4% | 31.7% |
| [GPT-5.4](https://benchgap.net/model/gpt-5-4.md) | 78.6% | 49.0% |
| [GPT-5.4 mini](https://benchgap.net/model/gpt-5-4-mini.md) | 51.2% | 9.8% |
| [GPT-5.4 nano](https://benchgap.net/model/gpt-5-4-nano.md) | 44.9% | 12.2% |
| [GPT-5.4 Pro](https://benchgap.net/model/gpt-5-4-pro.md) | 82.5% | 58.5% |
| [GPT-5.5](https://benchgap.net/model/gpt-5-5.md) | 85.3% | 72.5% |
| [GPT-5.5 Pro](https://benchgap.net/model/gpt-5-5-pro.md) | 87.7% | 78.0% |
| [GPT-5.6 Luna](https://benchgap.net/model/gpt-5-6-luna.md) | 78.6% | 58.5% |
| [GPT-5.6 Sol](https://benchgap.net/model/gpt-5-6-sol.md) | 89.0% | 83.0% |
| [GPT-5.6 Terra](https://benchgap.net/model/gpt-5-6-terra.md) | 84.9% | 68.3% |
| [GPT-6.1 Sol](https://benchgap.net/model/gpt-6-1-sol.md) | 93.7% | 100.0% |
| [GPT-6.1 Terra](https://benchgap.net/model/gpt-6-1-terra.md) | 90.9% | 92.7% |
| [GPT-6.2 Astra](https://benchgap.net/model/gpt-6-2-astra.md) | 92.6% | 100.0% |
| [GPT-6 Astra](https://benchgap.net/model/gpt-6-astra.md) | 93.7% | 97.6% |
| [GPT-6 Luna](https://benchgap.net/model/gpt-6-luna.md) | 78.9% | 56.1% |
| [GPT-6 Sol](https://benchgap.net/model/gpt-6-sol.md) | 89.8% | 90.0% |
| [Grok 4.3](https://benchgap.net/model/grok-4-3.md) | 42.8% | 14.6% |
| [Grok 4.5](https://benchgap.net/model/grok-4-5.md) | 57.2% | 24.4% |
| [Grok 4.6](https://benchgap.net/model/grok-4-6.md) | 66.0% | 31.7% |
| [Grok 4.7](https://benchgap.net/model/grok-4-7.md) | 53.0% | 17.1% |
| [Inkling](https://benchgap.net/model/inkling.md) | 33.3% | 4.9% |
| [Inkling-Small](https://benchgap.net/model/inkling-small.md) | 46.3% | 17.1% |
| [Kimi K2.6](https://benchgap.net/model/kimi-2-6.md) | 57.2% | 25.6% |
| [Kimi K2.7 Code](https://benchgap.net/model/kimi-k2-7-code.md) | 54.0% | 12.2% |
| [Muse Spark 1.3](https://benchgap.net/model/muse-spark-1-3.md) | 74.0% | 46.3% |
| [o4-mini](https://benchgap.net/model/o4-mini.md) | 36.1% | 4.9% |
| [Qwen3.7 Max](https://benchgap.net/model/qwen3-7-max.md) | 64.6% | 34.1% |
| [Qwen3.8 Max](https://benchgap.net/model/qwen3-8-max.md) | 74.7% | 46.3% |
