Benchmark Almanac

EBR-bench

GPT-6.1 Terra 78.1% leads since Oct 2026 Not in model cards

Agent plays the board game Earthborne Rangers 10-30 times in a row. Scored on objectives completed (of 21) to test learning from experience.

One card was banned in 2026-10 after it broke the game. Epoch expects saturation soon. Game material is not distributed.

Made by Epoch AI Released Jul 2026 Puzzles and chess Tasks kept private Last model added Oct 2026 Scores refreshed 31 Oct 2026
Level
Everyday Most adults, without training
Scored
Checked automatically A program checks the result against an answer key, unit tests or the expected final state.
Answers
Agent task Many steps in a terminal, browser, desktop, app, game or robot.
Paper Leaderboard
28
Models scored in Epoch AI data

GPT-6.1 Terra leads with 78.1%

Best score per model in Epoch AI’s data. Where a model ran with several settings or harnesses, we keep its best.

28 models. Scroll sideways for the rest.

  1. 78.1 GPT-6.1 Terra
  2. 76.2 GPT-6 Astra
  3. 71.4 Claude Opus 5.5
  4. 64.8 Gemini 4 Argon
  5. 57.1 Claude Fable 5.1
  6. 54.3 GPT-6.1 Sol
  7. 53.3 GPT-6 Sol
  8. 45.7 Claude Opus 5
  9. 44.8 GPT-5.6 Sol
  10. 43.8 Claude Haiku 5.5
  11. 39.5 Claude Fable 5
  12. 34.3 GPT-5.5
  13. 31.4 Muse Wonder
  14. 30.5 Grok 4.6
  15. 28.6 Claude Opus 4.8
  16. 25.4 GPT-5.4
  17. 23.0 GPT-5.2
  18. 19.1 Claude Opus 4.7
  19. 14.3 Claude Opus 4.5
  20. 14.3 Gemini 3.1 Pro
  21. 12.7 GPT-5
  22. 12.7 Claude Opus 4.6
  23. 9.52 Qwen3.7-Max
  24. 9.52 GLM-5.2
  25. 7.94 Claude Opus 4.1
  26. 4.76 Gemini 3.5 Flash
  27. 2.38 Claude Sonnet 4.5
  28. 2.38 Kimi K2.6
  1. 1 GPT-6.1 Terra max 78.1%
  2. 2 GPT-6 Astra max 76.2%
  3. 3 Claude Opus 5.5 max 71.4%
  4. 4 Gemini 4 Argon high 64.8%
  5. 5 Claude Fable 5.1 max 57.1%
  6. 6 GPT-6.1 Sol max 54.3%
  7. 7 GPT-6 Sol max 53.3%
  8. 8 Claude Opus 5 max 45.7%
  9. 9 GPT-5.6 Sol max 44.8%
  10. 10 Claude Haiku 5.5 max 43.8%
  11. 11 Claude Fable 5 max 39.5%
  12. 12 GPT-5.5 xhigh 34.3%
  13. 13 Muse Wonder max 31.4%
  14. 14 Grok 4.6 xhigh 30.5%
  15. 15 Claude Opus 4.8 max 28.6%
  16. 16 GPT-5.4 xhigh 25.4%
  17. 17 GPT-5.2 xhigh 23%
  18. 18 Claude Opus 4.7 max 19.1%
  19. 19 Claude Opus 4.5 128k thinking 14.3%
  20. 20 Gemini 3.1 Pro 14.3%
  21. 21 GPT-5 high 12.7%
  22. 22 Claude Opus 4.6 max 12.7%
  23. 23 Qwen3.7-Max 9.52%
  24. 24 GLM-5.2 max 9.52%
  25. 25 Claude Opus 4.1 7.94%
  26. 26 Gemini 3.5 Flash high 4.76%
  27. 27 Claude Sonnet 4.5 no thinking 2.38%
  28. 28 Kimi K2.6 2.38%
#ModelScoreSettingDateSource
1 GPT-6.1 Terra 78.1% max Oct 2026 Epoch AI
2 GPT-6 Astra 76.2% max Sep 2026 Epoch AI
3 Claude Opus 5.5 71.4% max Sep 2026 Epoch AI
4 Gemini 4 Argon 64.8% high Sep 2026 Epoch AI
5 Claude Fable 5.1 57.1% max Sep 2026 Epoch AI
6 GPT-6.1 Sol 54.3% max Sep 2026 Epoch AI
7 GPT-6 Sol 53.3% max Sep 2026 Epoch AI
8 Claude Opus 5 45.7% max Jul 2026 Epoch AI
9 GPT-5.6 Sol 44.8% max Jul 2026 Epoch AI
10 Claude Haiku 5.5 43.8% max Oct 2026 Epoch AI
11 Claude Fable 5 39.5% max Jun 2026 Epoch AI
12 GPT-5.5 34.3% xhigh Apr 2026 Epoch AI
13 Muse Wonder 31.4% max Oct 2026 Epoch AI
14 Grok 4.6 30.5% xhigh Aug 2026 Epoch AI
15 Claude Opus 4.8 28.6% max May 2026 Epoch AI
16 GPT-5.4 25.4% xhigh Mar 2026 Epoch AI
17 GPT-5.2 23% xhigh Dec 2025 Epoch AI
18 Claude Opus 4.7 19.1% max Apr 2026 Epoch AI
19 Claude Opus 4.5 14.3% 128k thinking Nov 2025 Epoch AI
20 Gemini 3.1 Pro 14.3% Feb 2026 Epoch AI
21 GPT-5 12.7% high Aug 2025 Epoch AI
22 Claude Opus 4.6 12.7% max Feb 2026 Epoch AI
23 Qwen3.7-Max 9.52% May 2026 Epoch AI
24 GLM-5.2 9.52% max Jun 2026 Epoch AI
25 Claude Opus 4.1 7.94% Aug 2025 Epoch AI
26 Gemini 3.5 Flash 4.76% high May 2026 Epoch AI
27 Claude Sonnet 4.5 2.38% no thinking Sep 2025 Epoch AI
28 Kimi K2.6 2.38% Apr 2026 Epoch AI

6 models have held number one since its release in Jul 2026

Lab claims count from the day the card came out, independent runs from the model’s release.

Anthropic OpenAI
Epoch AI runs 6 leaders
GPT-6.1 Terra 78.1%

Scores over time

All the scores we have, one dot each, so a model tested at several settings appears more than once.

Run or collected by Epoch AI Best so far
0% 20% 40% 60% 80% 100% 2026 2027 Near the ceiling (85%) GPT-6.1 Terra Claude Fable 5.1 Claude Opus 5 Claude Fable 5 GPT-5.5 GPT-5.4 Claude Opus 4.5 0% 20% 40% 60% 80% 100% 2026 2027 Near the ceiling (85%) GPT-6.1 Terra Claude Fable 5.1

No lab model card reports it yet

We read the headline results of 479 launch posts from 51 labs, and none of them list this benchmark.

More in Games and long tasks

All 68 in Games and long tasks
  • ARC-AGI-3 In use Interactive game environments where agents must explore, infer goals and rules, and act efficiently without instructions. GPT-6 Astra 100%
  • MineDojo Fading Agents complete open-ended tasks in Minecraft, such as mining blocks and fighting mobs, from a large simulation suite. Programmatic checks score many tasks. Seed2.0 Pro 49%
  • Orak Fading LLM agents play 12 popular video games across most genres through an MCP interface. Scores come from in-game progress or outcomes per title. Nemotron 3 Nano 4B 22.9
  • GameWorld Only in system cards Multimodal game agents play 34 browser games across 170 tasks. Outcomes are checked from the game state. Gemini 3 Flash 41.9%
  • Swarm Bench Only in system cards Models orchestrate agent swarms on complex tasks that benefit from coordinated decomposition and parallel execution. Kimi K3 76.3%
  • TextQuests Only in system cards Agents play 25 classic interactive fiction games. Measures game progress and harmful actions per 100 steps. Muse Spark 2.40