# FrontierMath v2 (Tier 4) leaderboard

Mathematics benchmark

> Every figure below is reproduced as its upstream source published it: nothing is modelled, estimated, interpolated or converted. `Unavailable` means no source published the value — it is never a zero. Each value carries its evidence state and the date it was observed.

Page: https://modelscale.dev/leaderboards/benchmarks/math-frontier-math-v2-tier4

FrontierMath v2 Tier 4. Every model the catalog carries a published FrontierMath v2 (Tier 4) value for, ranked by that value.

## Benchmark record

| Field | Value |
| --- | --- |
| Category | Mathematics |
| Measure | Python-enabled iterative mathematical problem solving |
| Tasks | 41 private extreme-difficulty mathematics problems |
| Difficulty | Research-level mathematics requiring hours or days of expert work |
| Published by | [FrontierMath Tier 4 v2 leaderboard](https://epoch.ai/benchmarks/frontiermath-tier-4-v2?view=graph&tab=leaderboard) |

Epoch AI's corrected v2 Tier 4 expansion, a separate set of exceptionally difficult research-level mathematics problems evaluated with Python-enabled iterative reasoning.

## FrontierMath v2 (Tier 4) ranking

81 models with a published FrontierMath v2 (Tier 4) value, ordered by that value, highest first. Models the source has not scored on this benchmark are not listed — they are unmeasured, not last.

| Rank | Model | Provider | Python-enabled iterative mathematical problem solving | Evidence |
| --- | --- | --- | --- | --- |
| 1 | [Claude Fable 5.2](https://modelscale.dev/models/claude-fable-5-2) | Anthropic | 100.000 | Observed 2026-11-21 · source benchlm:models |
| 1 | [GPT-6.1 Sol](https://modelscale.dev/models/gpt-6-1-sol) | OpenAI | 100.000 | Observed 2026-11-21 · source benchlm:models |
| 1 | [GPT-6.2 Astra](https://modelscale.dev/models/gpt-6-2-astra) | OpenAI | 100.000 | Observed 2026-11-21 · source benchlm:models |
| 4 | [GPT-6 Astra](https://modelscale.dev/models/gpt-6-astra) | OpenAI | 97.600 | Observed 2026-11-21 · source benchlm:models |
| 5 | [Gemini 4 Carbon](https://modelscale.dev/models/gemini-4-carbon) | Google | 95.122 | Observed 2026-11-21 · source benchlm:models |
| 6 | [Claude Opus 5.5](https://modelscale.dev/models/claude-opus-5-5) | Anthropic | 95.000 | Observed 2026-11-21 · source benchlm:models |
| 7 | [GPT-6.1 Terra](https://modelscale.dev/models/gpt-6-1-terra) | OpenAI | 92.683 | Observed 2026-11-21 · source benchlm:models |
| 8 | [Claude Fable 5](https://modelscale.dev/models/claude-fable-5) | Anthropic | 90.200 | Observed 2026-11-21 · source benchlm:models |
| 9 | [GPT-6 Sol](https://modelscale.dev/models/gpt-6-sol) | OpenAI | 90.000 | Observed 2026-11-21 · source benchlm:models |
| 10 | [Claude Fable 5.1](https://modelscale.dev/models/claude-fable-5-1) | Anthropic | 87.805 | Observed 2026-11-21 · source benchlm:models |
| 10 | [Gemini 4 Hydrogen](https://modelscale.dev/models/gemini-4-hydrogen) | Google | 87.805 | Observed 2026-11-21 · source benchlm:models |
| 12 | [GPT-5.6 Sol](https://modelscale.dev/models/gpt-5-6-sol) | OpenAI | 83.000 | Observed 2026-11-21 · source benchlm:models |
| 13 | [Claude Sonnet 5.5](https://modelscale.dev/models/claude-sonnet-5-5) | Anthropic | 80.488 | Observed 2026-11-21 · source benchlm:models |
| 14 | [Gemini 4 Argon](https://modelscale.dev/models/gemini-4-argon) | Google | 78.049 | Observed 2026-11-21 · source benchlm:models |
| 14 | [GPT-5.5 Pro](https://modelscale.dev/models/gpt-5-5-pro) | OpenAI | 78.049 | Observed 2026-11-21 · source benchlm:models |
| 16 | [Claude Opus 5](https://modelscale.dev/models/claude-opus-5) | Anthropic | 73.171 | Observed 2026-11-21 · source benchlm:models |
| 17 | [GPT-5.5](https://modelscale.dev/models/gpt-5-5) | OpenAI | 72.500 | Observed 2026-11-21 · source benchlm:models |
| 18 | [GPT-5.6 Terra](https://modelscale.dev/models/gpt-5-6-terra) | OpenAI | 68.300 | Observed 2026-11-21 · source benchlm:models |
| 19 | [GPT-5.4 Pro](https://modelscale.dev/models/gpt-5-4-pro) | OpenAI | 58.537 | Observed 2026-11-21 · source benchlm:models |
| 20 | [GPT-5.6 Luna](https://modelscale.dev/models/gpt-5-6-luna) | OpenAI | 58.500 | Observed 2026-11-21 · source benchlm:models |
| 21 | [Claude Opus 4.8](https://modelscale.dev/models/claude-opus-4-8) | Anthropic | 56.098 | Observed 2026-11-21 · source benchlm:models |
| 21 | [GPT-6 Luna](https://modelscale.dev/models/gpt-6-luna) | OpenAI | 56.098 | Observed 2026-11-21 · source benchlm:models |
| 23 | [GPT-5.4](https://modelscale.dev/models/gpt-5-4) | OpenAI | 49.000 | Observed 2026-11-21 · source benchlm:models |
| 24 | [Claude Haiku 5.5](https://modelscale.dev/models/claude-haiku-5-5) | Anthropic | 46.341 | Observed 2026-11-21 · source benchlm:models |
| 24 | [Muse Spark 1.3](https://modelscale.dev/models/muse-spark-1-3) | Meta | 46.341 | Observed 2026-11-21 · source benchlm:models |
| 24 | [Qwen3.8 Max](https://modelscale.dev/models/qwen3-8-max) | Alibaba | 46.341 | Observed 2026-11-21 · source benchlm:models |
| 27 | [Gemini 3.7 Flash](https://modelscale.dev/models/gemini-3-7-flash) | Google | 36.585 | Observed 2026-11-21 · source benchlm:models |
| 28 | [Qwen3.7 Max](https://modelscale.dev/models/qwen3-7-max) | Alibaba | 34.146 | Observed 2026-11-21 · source benchlm:models |
| 29 | [Claude Opus 4.7](https://modelscale.dev/models/claude-opus-4-7) | Anthropic | 31.707 | Observed 2026-11-21 · source benchlm:models |
| 29 | [Grok 4.6](https://modelscale.dev/models/grok-4-6) | xAI | 31.707 | Observed 2026-11-21 · source benchlm:models |
| 31 | [GPT-5.2](https://modelscale.dev/models/gpt-5-2) | OpenAI | 31.700 | Observed 2026-11-21 · source benchlm:models |
| 32 | [Claude Sonnet 5](https://modelscale.dev/models/claude-sonnet-5) | Anthropic | 29.268 | Observed 2026-11-21 · source benchlm:models |
| 32 | [GLM-5.2](https://modelscale.dev/models/glm-5-2) | Z.AI | 29.268 | Observed 2026-11-21 · source benchlm:models |
| 32 | [GLM-5.3](https://modelscale.dev/models/glm-5-3) | Z.AI | 29.268 | Observed 2026-11-21 · source benchlm:models |
| 35 | [Claude Opus 4.6](https://modelscale.dev/models/claude-opus-4-6) | Anthropic | 26.829 | Observed 2026-11-21 · source benchlm:models |
| 35 | [DeepSeek V4 Pro](https://modelscale.dev/models/deepseek-v4-pro) | DeepSeek | 26.829 | Observed 2026-11-21 · source benchlm:models |
| 35 | [DeepSeek V4.1 Flash 1017](https://modelscale.dev/models/deepseek-v4-1-flash) | DeepSeek | 26.829 | Observed 2026-11-21 · source benchlm:models |
| 35 | [Gemini 3.1 Pro](https://modelscale.dev/models/gemini-3-1-pro) | Google | 26.829 | Observed 2026-11-21 · source benchlm:models |
| 35 | [Gemini 3.5 Flash](https://modelscale.dev/models/gemini-3-5-flash) | Google | 26.829 | Observed 2026-11-21 · source benchlm:models |
| 40 | [Kimi K2.6](https://modelscale.dev/models/kimi-2-6) | Moonshot AI | 25.641 | Observed 2026-11-21 · source benchlm:models |
| 41 | [Grok 4.5](https://modelscale.dev/models/grok-4-5) | xAI | 24.390 | Observed 2026-11-21 · source benchlm:models |
| 42 | [Gemini 3.6 Flash](https://modelscale.dev/models/gemini-3-6-flash) | Google | 21.951 | Observed 2026-11-21 · source benchlm:models |
| 42 | [Gemini 3.8 Flash](https://modelscale.dev/models/gemini-3-8-flash) | Google | 21.951 | Observed 2026-11-21 · source benchlm:models |
| 44 | [Gemini 3 Pro](https://modelscale.dev/models/gemini-3-pro) | Google | 18.750 | Last good 2026-10-08 · source benchlm:models |
| 45 | [Gemini 3 Flash](https://modelscale.dev/models/gemini-3-flash) | Google | 17.073 | Observed 2026-11-21 · source benchlm:models |
| 45 | [GLM-5.3-Flash](https://modelscale.dev/models/glm-5-3-flash) | Z.AI | 17.073 | Observed 2026-11-21 · source benchlm:models |
| 45 | [Grok 4.7](https://modelscale.dev/models/grok-4-7) | xAI | 17.073 | Observed 2026-11-21 · source benchlm:models |
| 45 | [Inkling-Small](https://modelscale.dev/models/inkling-small) | Thinking Machines Lab | 17.073 | Observed 2026-11-21 · source benchlm:models |
| 49 | [Grok 4.3](https://modelscale.dev/models/grok-4-3) | xAI | 14.634 | Observed 2026-11-21 · source benchlm:models |
| 50 | [Muse Spark](https://modelscale.dev/models/muse-spark) | Meta | 14.600 | Last good 2026-10-08 · source benchlm:models |
| 51 | [GLM-5.1](https://modelscale.dev/models/glm-5-1) | Z.AI | 12.500 | Last good 2026-10-08 · source benchlm:models |
| 51 | [GPT-5.1](https://modelscale.dev/models/gpt-5-1) | OpenAI | 12.500 | Last good 2026-10-08 · source benchlm:models |
| 53 | [GPT-5.4 nano](https://modelscale.dev/models/gpt-5-4-nano) | OpenAI | 12.195 | Observed 2026-11-21 · source benchlm:models |
| 53 | [Kimi K2.7 Code](https://modelscale.dev/models/kimi-k2-7-code) | Moonshot AI | 12.195 | Observed 2026-11-21 · source benchlm:models |
| 55 | [GPT-5.4 mini](https://modelscale.dev/models/gpt-5-4-mini) | OpenAI | 9.756 | Observed 2026-11-21 · source benchlm:models |
| 56 | [Qwen3.6 Plus](https://modelscale.dev/models/qwen3-6-plus) | Alibaba | 8.333 | Last good 2026-10-08 · source benchlm:models |
| 57 | [Claude Sonnet 4.6](https://modelscale.dev/models/claude-sonnet-4-6) | Anthropic | 8.300 | Last good 2026-10-08 · source benchlm:models |
| 58 | [Claude Opus 4.5](https://modelscale.dev/models/claude-opus-4-5) | Anthropic | 4.878 | Observed 2026-11-21 · source benchlm:models |
| 58 | [Inkling](https://modelscale.dev/models/inkling) | Thinking Machines Lab | 4.878 | Observed 2026-11-21 · source benchlm:models |
| 58 | [o4-mini](https://modelscale.dev/models/o4-mini) | OpenAI | 4.878 | Observed 2026-11-21 · source benchlm:models |
| 58 | [o4-mini (high)](https://modelscale.dev/models/o4-mini-high) | OpenAI | 4.878 | Last good 2026-11-18 · source benchlm:models |
| 62 | [Kimi K2.5](https://modelscale.dev/models/kimi-k2-5) | Moonshot AI | 4.200 | Last good 2026-10-08 · source benchlm:models |
| 63 | [Gemini 2.5 Flash](https://modelscale.dev/models/gemini-2-5-flash) | Google | 4.167 | Last good 2026-10-08 · source benchlm:models |
| 63 | [Qwen 3.6 Max (preview)](https://modelscale.dev/models/qwen3-6-max-preview) | Alibaba | 4.167 | Last good 2026-10-08 · source benchlm:models |
| 65 | [Claude Sonnet 4.5](https://modelscale.dev/models/claude-sonnet-4-5) | Anthropic | 2.439 | Observed 2026-11-21 · source benchlm:models |
| 66 | [GLM-4.6](https://modelscale.dev/models/glm-4-6) | Z.AI | 2.128 | Last good 2026-10-08 · source benchlm:models |
| 67 | [DeepSeek V3.2](https://modelscale.dev/models/deepseek-v3-2) | DeepSeek | 2.100 | Last good 2026-10-08 · source benchlm:models |
| 67 | [GLM-5](https://modelscale.dev/models/glm-5) | Z.AI | 2.100 | Last good 2026-10-08 · source benchlm:models |
| 69 | [Claude Haiku 4.5](https://modelscale.dev/models/claude-haiku-4-5) | Anthropic | 2.083 | Last good 2026-10-08 · source benchlm:models |
| 69 | [Grok 4](https://modelscale.dev/models/grok-4) | xAI | 2.083 | Last good 2026-10-08 · source benchlm:models |
| 69 | [o3](https://modelscale.dev/models/o3) | OpenAI | 2.083 | Last good 2026-10-08 · source benchlm:models |
| 69 | [Qwen3.5 Plus](https://modelscale.dev/models/qwen3-5-plus) | Alibaba | 2.083 | Last good 2026-10-08 · source benchlm:models |
| 73 | [Claude 3.5 Sonnet](https://modelscale.dev/models/claude-3-5-sonnet) | Anthropic | 0.000 | Last good 2026-10-08 · source benchlm:models |
| 73 | [Gemini 2.5 Pro](https://modelscale.dev/models/gemini-2-5-pro) | Google | 0.000 | Observed 2026-11-21 · source benchlm:models |
| 73 | [Gemini 3.5 Flash-Lite](https://modelscale.dev/models/gemini-3-5-flash-lite) | Google | 0.000 | Observed 2026-11-21 · source benchlm:models |
| 73 | [GLM-4.7](https://modelscale.dev/models/glm-4-7) | Z.AI | 0.000 | Last good 2026-10-08 · source benchlm:models |
| 73 | [GPT-4.1](https://modelscale.dev/models/gpt-4-1) | OpenAI | 0.000 | Last good 2026-10-08 · source benchlm:models |
| 73 | [Grok 3 [Beta]](https://modelscale.dev/models/grok-3-beta) | xAI | 0.000 | Last good 2026-10-08 · source benchlm:models |
| 73 | [Kimi K2](https://modelscale.dev/models/kimi-k2) | Moonshot AI | 0.000 | Last good 2026-10-08 · source benchlm:models |
| 73 | [Qwen3 235B 2507 (Reasoning)](https://modelscale.dev/models/qwen3-235b-2507-reasoning) | Alibaba | 0.000 | Last good 2026-10-08 · source benchlm:models |
| 73 | [Qwen3.5 Flash](https://modelscale.dev/models/qwen3-5-flash) | Alibaba | 0.000 | Last good 2026-10-08 · source benchlm:models |

Rows are ordered by the value the source published, highest first. The source does not state whether a higher value is the better result, so this document does not either: for a benchmark that measures a rate of failure, read the table from the bottom.
