Benchmark Almanac

Olam Labs Multi-Agent Arena

GPT-6 Astra 1,544 leads since September 2026
Not in model cards

People and anonymized models share a table in social strategy games: Diplomacy, Social Poker, Codenames and Risk. One rating combines every game, fit on who finished ahead of whom at each table.

Every model plays through the same Pi harness, under a generic player name, so no seat knows which players are people. The evaluations page shows the top 16 models of each chart. By October 14th, 2026 the rating covered four games.

Made by
Olam Labs
Released
July 2026
Kind
Multi-agent games
Level
Everyday
Scored
Ranked by matches
Answers
Agent task
Tasks
Kept private
Measures
Capability
Language
English
In and out
Text → Text, Actions
Interaction
Multi-agent
Items from
Live play or votes
Refresh
Live
Run by
A host
In views
Games, Agents, Kept fresh
Strongest model tested
GPT-6 Astra #4 on the capability index
Last model added
October 2026
Scores refreshed
November 20th, 2026
16
Models scored on its own leaderboard

GPT-6 Astra leads with 1,544

Best score per model from the benchmark's own leaderboard, read on November 20th, 2026. Where a model ran with several settings or harnesses, we keep its best.

Bars start at 1,490.

  1. 1 GPT-6 Astra 1,544
  2. 2 GPT-6.1 Sol 1,540
  3. 3 Claude Fable 5 1,531
  4. 4 GPT-6 Sol 1,523
  5. 5 Gemini 4 Argon 1,523
  6. 6 GPT-5.5 1,521
  7. 7 GPT-5.6 Sol 1,517
  8. 8 Claude Fable 5.1 1,517
  9. 9 Claude Opus 5 1,514
  10. 10 Claude Opus 5.5 1,513
  11. 11 Gemini 3.8 Flash 1,509
  12. 12 Grok 4.7 1,508
  13. 13 Claude Sonnet 5.5 1,506
  14. 14 Muse Wonder 1,506
  15. 15 Grok 4.6 1,498
  16. 16 Kimi K3 1,496

Source: Official Olam Labs Multi-Agent Arena leaderboard

#ModelScoreSettingReleasedRun onSource
1 GPT-6 Astra 1,544 Pi Sep 2026 Oct 2026 Official leaderboard
2 GPT-6.1 Sol 1,540 Pi Sep 2026 Oct 2026 Official leaderboard
3 Claude Fable 5 1,531 Pi Jun 2026 Oct 2026 Official leaderboard
4 GPT-6 Sol 1,523 Pi Sep 2026 Oct 2026 Official leaderboard
5 Gemini 4 Argon 1,523 Pi Sep 2026 Oct 2026 Official leaderboard
6 GPT-5.5 1,521 Pi Apr 2026 Oct 2026 Official leaderboard
7 GPT-5.6 Sol 1,517 Pi Jul 2026 Oct 2026 Official leaderboard
8 Claude Fable 5.1 1,517 Pi Sep 2026 Oct 2026 Official leaderboard
9 Claude Opus 5 1,514 Pi Jul 2026 Oct 2026 Official leaderboard
10 Claude Opus 5.5 1,513 Pi Sep 2026 Oct 2026 Official leaderboard
11 Gemini 3.8 Flash 1,509 Pi Sep 2026 Oct 2026 Official leaderboard
12 Grok 4.7 1,508 Pi Sep 2026 Oct 2026 Official leaderboard
13 Claude Sonnet 5.5 1,506 Pi Sep 2026 Oct 2026 Official leaderboard
14 Muse Wonder 1,506 Pi Oct 2026 Oct 2026 Official leaderboard
15 Grok 4.6 1,498 Pi Aug 2026 Oct 2026 Official leaderboard
16 Kimi K3 1,496 Pi Jul 2026 Oct 2026 Official leaderboard

2 models have held number one on the official leaderboard since its release in July 2026

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

Anthropic OpenAI
Own leaderboard 2 leaders
GPT-6 Astra 1,544

Source: Official Olam Labs Multi-Agent Arena leaderboard

Scores over time

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

Plot with other benchmarks
On the benchmark’s own leaderboard
1,500 1,520 1,540 2026 1,500 1,520 1,540 2026

Source: Official Olam Labs Multi-Agent Arena leaderboard

No lab model card reports it yet

We read the headline results of 540 launch posts from 57 labs, and none of them list this benchmark.

More in Games and long tasks

All 261 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-6.2 Astra 100%
  • Orak Fading LLM agents play 12 popular video games across most genres through an MCP interface. Each game scores in-game progress or outcomes, and the board ranks models by their average rank across the games. Gemini 2.5 Pro 3.90
  • 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%
  • TextQuests Only in system cards Agents play 25 classic interactive fiction games. The score is harmful actions per 100 steps, where lower is better. The board also reports game progress. Llama 4 Scout 7.00
  • GameWorld Only in system cards Multimodal game agents play 34 browser games across 170 tasks. Outcomes are checked from the game state. GPT-6.1 Terra 76.8%
  • 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%

Data audit

All 10 checks pass. Every benchmark

Catalog record 2 pass
  • Pass Task access The tasks are private.
  • Pass Confidence Checked against a primary source.
Scores 4 pass 1 note
  • Pass Independent runs Scores from the official leaderboard (16 entries).
  • Pass Models scored 16 models we track.
  • Pass Newest score The newest is from October 2026, for Muse Wonder.
  • Pass Frontier models 6 of the top 10 have a score, led by GPT-6 Astra (#4).
  • Note Headroom Scored as a rating, with no fixed ceiling.
Leaderboard import 3 pass
  • Pass Importer We read the board on olamlabs.ai: 16 rows for models we track at the last read.
  • Pass Last read November 20th, 2026.
  • Pass Read worked The last read worked.
Consistency 1 pass 1 note
  • Note Lab claims No model has both a lab number and an independent run.
  • Pass Known issues No published audit found problems.

How we check