# Agents' Last Exam leaderboard

Agentic 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://modelchoice.dev/leaderboards/benchmarks/agentic-agents-last-exam

Every model the catalog carries a published Agents' Last Exam value for, ranked by that value.

## Benchmark record

| Field | Value |
| --- | --- |
| Category | Agentic |
| Measure | Provider-reported task score |
| Tasks | Agent tasks |
| Difficulty | Advanced agentic work |
| Published by | [DeepSeek V4 Flash 0731 update](https://api-docs.deepseek.com/zh-cn/updates/) |

An agent benchmark reported in DeepSeek's V4 Flash 0731 launch comparison.

## Agents' Last Exam ranking

22 models with a published Agents' Last Exam 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 | Provider-reported task score | Evidence |
| --- | --- | --- | --- | --- |
| 1 | [GPT-6.1 Terra](https://modelchoice.dev/models/gpt-6-1-terra) | OpenAI | 60.0 | Observed 2026-11-21 · source benchlm:models |
| 2 | [GPT-6 Astra](https://modelchoice.dev/models/gpt-6-astra) | OpenAI | 59.3 | Observed 2026-11-21 · source benchlm:models |
| 3 | [GPT-6 Sol](https://modelchoice.dev/models/gpt-6-sol) | OpenAI | 56.4 | Observed 2026-11-21 · source benchlm:models |
| 4 | [Qwen3.8 Max](https://modelchoice.dev/models/qwen3-8-max) | Alibaba | 52.4 | Observed 2026-11-21 · source benchlm:models |
| 5 | [Qwen3.8-Flash-Next](https://modelchoice.dev/models/qwen3-8-flash-next) | Alibaba | 51.2 | Observed 2026-11-21 · source benchlm:models |
| 6 | [Kimi K3.1 Pro](https://modelchoice.dev/models/kimi-3-1-pro) | Moonshot AI | 43.2 | Observed 2026-11-21 · source benchlm:models |
| 7 | [Qwen3.8-27B](https://modelchoice.dev/models/qwen3-8-27b) | Alibaba | 42.9 | Observed 2026-11-21 · source benchlm:models |
| 8 | [Hy4](https://modelchoice.dev/models/hy4) | Tencent | 41.0 | Observed 2026-11-21 · source benchlm:models |
| 9 | [Gemini 4 Argon](https://modelchoice.dev/models/gemini-4-argon) | Google | 39.5 | Observed 2026-11-21 · source benchlm:models |
| 10 | [LongCat-2.5](https://modelchoice.dev/models/longcat-2-5) | Meituan | 38.5 | Observed 2026-11-21 · source benchlm:models |
| 11 | [DeepSeek V4.1 Flash 1017](https://modelchoice.dev/models/deepseek-v4-1-flash) | DeepSeek | 35.4 | Observed 2026-11-21 · source benchlm:models |
| 12 | [MiMo-V2.6-Pro](https://modelchoice.dev/models/mimo-v2-6-pro) | Xiaomi | 31.6 | Observed 2026-11-21 · source benchlm:models |
| 13 | [MiniMax M3.1 Flash](https://modelchoice.dev/models/minimax-m3-1-flash) | MiniMax | 29.6 | Observed 2026-11-21 · source benchlm:models |
| 14 | [Step 5 Preview](https://modelchoice.dev/models/step-5-preview) | StepFun | 29.5 | Observed 2026-11-21 · source benchlm:models |
| 15 | [GLM-5.3](https://modelchoice.dev/models/glm-5-3) | Z.AI | 28.5 | Observed 2026-11-21 · source benchlm:models |
| 16 | [MiMo-V2.6-Flash](https://modelchoice.dev/models/mimo-v2-6-flash) | Xiaomi | 27.6 | Observed 2026-11-21 · source benchlm:models |
| 17 | [Gemini 3.7 Flash](https://modelchoice.dev/models/gemini-3-7-flash) | Google | 26.3 | Observed 2026-11-21 · source benchlm:models |
| 17 | [GLM-5.3-Flash](https://modelchoice.dev/models/glm-5-3-flash) | Z.AI | 26.3 | Observed 2026-11-21 · source benchlm:models |
| 19 | [DeepSeek V4 Pro](https://modelchoice.dev/models/deepseek-v4-pro) | DeepSeek | 25.7 | Observed 2026-11-21 · source benchlm:models |
| 19 | [DeepSeek V4 Pro 0813](https://modelchoice.dev/models/deepseek-v4-pro-max) | DeepSeek | 25.7 | Last good 2026-11-18 · source benchlm:models |
| 21 | [DeepSeek V4 Flash 0731](https://modelchoice.dev/models/deepseek-v4-flash-max) | DeepSeek | 25.2 | Observed 2026-11-21 · source benchlm:models |
| 22 | [Hy4 preview](https://modelchoice.dev/models/hy4-preview) | Tencent | 22.8 | Observed 2026-11-21 · 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.
