Benchmark Almanac

Multi-Agent Arena Codenames

GPT-6 Astra 1,535 leads since September 2026
Not in model cards

Two teams of anonymized models play the word game Codenames. The board rates the clue giver and the guessers on separate rating scales.

The board has no single Codenames rating, so we show the clue giver’s (handler) rating and rank both roles as parts. It was rated from 352 games among models only, as of September 25th, 2026.

Made by
Olam Labs
Released
September 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 → Actions, Text
Interaction
Multi-agent
Environment
Codenames
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
September 2026
Scores refreshed
November 20th, 2026
16
Models scored on its own leaderboard

GPT-6 Astra leads with 1,535 among models released by September 2026

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.

The newest model with a score here came out in September 2026, and nobody has tested a newer one. Among the 20 highest-ranked models on the capability index, these have no score yet:

Bars start at 1,460.

  1. 1 GPT-6 Astra 1,535
  2. 2 Claude Fable 5 1,526
  3. 3 GPT-5.6 Sol 1,526
  4. 4 Grok 4.7 1,525
  5. 5 GPT-6 Sol 1,525
  6. 6 Gemini 3.8 Flash 1,521
  7. 7 Claude Opus 5.5 1,520
  8. 8 Claude Fable 5.1 1,519
  9. 9 Grok 4.6 1,506
  10. 10 DeepSeek V4.1 Flash 1,506
  11. 11 Claude Opus 5 1,502
  12. 12 Kimi K3 1,501
  13. 13 Muse Spark 1.3 1,496
  14. 14 GPT-6 Luna 1,475
  15. 15 GPT-5.6 Terra 1,469
  16. 16 GLM-5.3-Flash 1,469

Source: Official Multi-Agent Arena Codenames leaderboard

#ModelScore ± 95% CISettingReleasedRun onSource
1 GPT-6 Astra 1,535 1516 to 1552 Pi Sep 2026 Sep 2026 Official leaderboard
2 Claude Fable 5 1,526 1505 to 1543 Pi Jun 2026 Sep 2026 Official leaderboard
3 GPT-5.6 Sol 1,526 1504 to 1545 Pi Jul 2026 Sep 2026 Official leaderboard
4 Grok 4.7 1,525 ±25.5 Pi Sep 2026 Sep 2026 Official leaderboard
5 GPT-6 Sol 1,525 1499 to 1548 Pi Sep 2026 Sep 2026 Official leaderboard
6 Gemini 3.8 Flash 1,521 1495 to 1539 Pi Sep 2026 Sep 2026 Official leaderboard
7 Claude Opus 5.5 1,520 1492 to 1545 Pi Sep 2026 Sep 2026 Official leaderboard
8 Claude Fable 5.1 1,519 1496 to 1540 Pi Sep 2026 Sep 2026 Official leaderboard
9 Grok 4.6 1,506 1476 to 1530 Pi Aug 2026 Sep 2026 Official leaderboard
10 DeepSeek V4.1 Flash 1,506 1473 to 1532 Pi Sep 2026 Sep 2026 Official leaderboard
11 Claude Opus 5 1,502 1477 to 1523 Pi Jul 2026 Sep 2026 Official leaderboard
12 Kimi K3 1,501 1468 to 1526 Pi Jul 2026 Sep 2026 Official leaderboard
13 Muse Spark 1.3 1,496 1465 to 1522 Pi Sep 2026 Sep 2026 Official leaderboard
14 GPT-6 Luna 1,475 1437 to 1499 Pi Sep 2026 Sep 2026 Official leaderboard
15 GPT-5.6 Terra 1,469 1433 to 1494 Pi Jul 2026 Sep 2026 Official leaderboard
16 GLM-5.3-Flash 1,469 1431 to 1495 Pi Aug 2026 Sep 2026 Official leaderboard

Rankings by part

We read 2 parts from the benchmark’s own leaderboard: Handler and Operative. Each ranks models on one part of the test. GPT-6 Astra leads Handler and Claude Fable 5.1 leads Operative. Arrows show how far a model moved from its overall rank among the models we track.

Handler

Gives the one-word clues. 16 models we track have a score here.

  1. 1 GPT-6 Astra 1,535
  2. 2 Claude Fable 5 1,526
  3. 3 GPT-5.6 Sol 1,526
  4. 4 Grok 4.7 1,525
  5. 5 GPT-6 Sol 1,525
  6. 6 Gemini 3.8 Flash 1,521
  7. 7 Claude Opus 5.5 1,520
  8. 8 Claude Fable 5.1 1,519
  9. 9 Grok 4.6 1,506
  10. 10 DeepSeek V4.1 Flash 1,506

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

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

Anthropic OpenAI
Own leaderboard Not tested since Sep 2026
GPT-6 Astra 1,535

Source: Official Multi-Agent Arena Codenames 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
1,460 1,480 1,500 1,520 1,540 2026 1,460 1,480 1,500 1,520 1,540 2026

Source: Official Multi-Agent Arena Codenames 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 September 2026, for Claude Opus 5.5.
  • Pass Frontier models 3 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