Benchmark Almanac

Balatro Bench (Michael Andrzejewski)

GPT-6 Astra 18.0 leads since September 2026
Not in model cards

Agents play one fixed Balatro seed with the Red Deck on White Stake. The maker ranks runs by ante reached, breaking ties by best hand.

The maker lists scores for eight player models. Some runs use prior journals, seed information, or coaching. Two models appear only in unranked rows.

Released
August 2026
Kind
Video games
Level
Everyday
Scored
Checked automatically
Answers
Agent task
Tasks
All public
Measures
Capability
Language
English
In and out
Text → Actions
Interaction
Agentic
Environment
Balatro
Items from
Generated
Refresh
One fixed set
Run by
Anyone
In views
Games, Agents
Strongest model tested
Claude Fable 5.2 #2 on the capability index
Last model added
November 2026
Scores refreshed
November 20th, 2026
9
Entries on its own leaderboard, 8 models we track

GPT-6 Astra leads with 18.0 GPT-6 Astra leads with 18.0

Best score per entry from the benchmark's own leaderboard, read on November 20th, 2026. Where a model ran on its own with several settings, we keep its best. Every run is in the next chart. Agents other companies built around a model get a pale shade of that model's lab color. Gray marks blends and entries not in our model list, such as fine-tunes and older models. None of these count toward leaders. Agents built around a model, such as Opus 3 planner (Claude Opus 5 + Opus 3), are their own entries. We don't count them as the model's score. This view shows only models we track, each at its best.

  1. 1 GPT-6 Astra coached by michael, seed-informed 18.0
  2. 2 Claude Fable 5.1 personal context 12.0
  3. 3 GPT-5.6 Sol seed-informed 11.0
  4. 4 Claude Opus 5 blind, cold 11.0
  5. 5 Claude Fable 5.2 blind, cold 11.0
  6. 6 Claude Opus 4.8 blind + own journal 10.0
  7. 7 Claude Fable 5 blind, cold 10.0
  8. 8 Opus 3 planner (Claude Opus 5 + Opus 3) seed-informed hybrid 10.0
  9. 9 Claude Opus 5.5 seed-informed 10.0

Source: Official Balatro Bench (Michael Andrzejewski) leaderboard

#EntryScoreSettingReleasedRun onLinkSource
1 GPT-6 Astra 18.0 coached by michael, seed-informed Sep 2026 Sep 2026 github.com Official leaderboard
2 Claude Fable 5.1 12.0 personal context Sep 2026 Sep 2026 github.com Official leaderboard
3 GPT-5.6 Sol 11.0 seed-informed Jul 2026 Aug 2026 github.com Official leaderboard
4 Claude Opus 5 11.0 blind, cold Jul 2026 Jul 2026 github.com Official leaderboard
5 Claude Fable 5.2 11.0 blind, cold Nov 2026 Nov 2026 github.com Official leaderboard
6 Claude Opus 4.8 10.0 blind + own journal May 2026 Jul 2026 github.com Official leaderboard
7 Claude Fable 5 10.0 blind, cold Jun 2026 Jul 2026 github.com Official leaderboard
8 Opus 3 planner (Claude Opus 5 + Opus 3) 10.0 seed-informed hybrid Jul 2026 Aug 2026 github.com Official leaderboard
9 Claude Opus 5.5 10.0 seed-informed Sep 2026 Sep 2026 github.com Official leaderboard
#ModelScoreSettingReleasedRun onLinkSource
1 GPT-6 Astra 18.0 coached by michael, seed-informed Sep 2026 Sep 2026 github.com Official leaderboard
2 Claude Fable 5.1 12.0 personal context Sep 2026 Sep 2026 github.com Official leaderboard
3 GPT-5.6 Sol 11.0 seed-informed Jul 2026 Aug 2026 github.com Official leaderboard
4 Claude Opus 5 11.0 blind, cold Jul 2026 Jul 2026 github.com Official leaderboard
5 Claude Fable 5.2 11.0 blind, cold Nov 2026 Nov 2026 github.com Official leaderboard
6 Claude Opus 4.8 10.0 blind + own journal May 2026 Jul 2026 github.com Official leaderboard
7 Claude Fable 5 10.0 blind, cold Jun 2026 Jul 2026 github.com Official leaderboard
8 Claude Opus 5.5 10.0 seed-informed Sep 2026 Sep 2026 github.com Official leaderboard

Scores at each effort level

Models the leaderboard ran at two or more effort levels. We don’t have enough cost data here to say what the extra effort costs.

Lowest effort to highest, light to dark
GPT-6 Astra
seed-inform… coached by …
18.0
Claude Fable 5.1
coached by … personal co…
12.0
GPT-5.6 Sol
blind, cold
11.0
Claude Opus 5
seed-only same-sessio… blind, cold
11.0
Claude Opus 4.8
guided blind, cold blind + own…
10.0
Claude Fable 5
seed-inform… blind, cold
10.0
Claude Opus 5.5
personal co…
10.0

Source: Official Balatro Bench (Michael Andrzejewski) leaderboard

ModelEffortHarnessScoreCost per taskOutput tokens
GPT-6 Astra coached by michael, seed-informed 18.0
GPT-6 Astra seed-informed 13.0
Claude Fable 5.1 personal context 12.0
GPT-5.6 Sol seed-informed 11.0
Claude Opus 5 blind, cold 11.0
Claude Fable 5.2 blind, cold 11.0
Claude Opus 4.8 blind + own journal 10.0
Claude Fable 5 blind, cold 10.0
GPT-5.6 Sol blind, cold 10.0
Claude Opus 5 coached by michael 10.0
Claude Opus 5 michael-principles, seed-informed hybrid Opus 5 planner 10.0
Claude Opus 5.5 seed-informed 10.0
Opus 3 planner (Claude Opus 5 + Opus 3) seed-informed hybrid Opus 3 planner 10.0
Claude Opus 5 same-session replay 9.00
Claude Opus 5 seed-informed 9.00
Claude Opus 5.5 personal context 9.00
Claude Opus 4.8 blind, cold 6.00
Claude Fable 5 seed-informed 6.00
Claude Opus 5 seed-only 5.00
Claude Fable 5.1 coached by Michael 4.00
Claude Opus 4.8 guided 3.00

Scores at each effort level

Models the leaderboard ran at two or more effort levels. We don’t have enough cost data here to say what the extra effort costs.

Lowest effort to highest, light to dark
GPT-6 Astra
seed-inform… coached by …
18.0
Claude Fable 5.1
coached by … personal co…
12.0
GPT-5.6 Sol
blind, cold
11.0
Claude Opus 5
seed-only same-sessio… blind, cold
11.0
Claude Opus 4.8
guided blind, cold blind + own…
10.0
Claude Fable 5
seed-inform… blind, cold
10.0
Claude Opus 5.5
personal co…
10.0

Source: Official Balatro Bench (Michael Andrzejewski) leaderboard

ModelEffortHarnessScoreCost per taskOutput tokens
GPT-6 Astra coached by michael, seed-informed 18.0
GPT-6 Astra seed-informed 13.0
Claude Fable 5.1 personal context 12.0
GPT-5.6 Sol seed-informed 11.0
Claude Opus 5 blind, cold 11.0
Claude Fable 5.2 blind, cold 11.0
Claude Opus 4.8 blind + own journal 10.0
Claude Fable 5 blind, cold 10.0
GPT-5.6 Sol blind, cold 10.0
Claude Opus 5 coached by michael 10.0
Claude Opus 5 michael-principles, seed-informed hybrid Opus 5 planner 10.0
Claude Opus 5.5 seed-informed 10.0
Claude Opus 5 same-session replay 9.00
Claude Opus 5 seed-informed 9.00
Claude Opus 5.5 personal context 9.00
Claude Opus 4.8 blind, cold 6.00
Claude Fable 5 seed-informed 6.00
Claude Opus 5 seed-only 5.00
Claude Fable 5.1 coached by Michael 4.00
Claude Opus 4.8 guided 3.00

3 models have held number one on the official leaderboard since its release in August 2026

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

OpenAI Anthropic
Own leaderboard 3 leaders

Source: Official Balatro Bench (Michael Andrzejewski) 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
0.00 5.00 10.0 15.0 20.0 2026 2027 0.00 5.00 10.0 15.0 20.0 2026 2027

Source: Official Balatro Bench (Michael Andrzejewski) 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 All tasks are public.
  • Pass Confidence Checked against a primary source.
Scores 4 pass 1 note
  • Pass Independent runs Scores from the official leaderboard (9 entries).
  • Pass Models scored 8 models we track, plus 1 other entry on its leaderboard.
  • Pass Newest score The newest is from November 2026, for Claude Fable 5.2.
  • Pass Frontier models 4 of the top 10 have a score, led by Claude Fable 5.2 (#2).
  • Note Headroom Scored as a plain number, with no fixed ceiling.
Leaderboard import 3 pass
  • Pass Importer We read the board on github.com: 21 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