Benchmark Almanac

Portal2Bench

GPT-6.2 Astra 26.0 leads since November 2026
Not in model cards

Agents solve Portal 2 chambers using screenshots, position readings, and game controls. Each chamber counts if the agent reaches the next map within $30 and 300 plans.

The leaderboard reports scores for nine models. Eight attempted all 26 ranked chambers. Fable 5.1 attempted 15.

Made by
Kingston Yip
Released
November 2026
Kind
Video games
Level
Everyday
Scored
Checked automatically
Answers
Agent task
Measures
Capability
Language
English
In and out
Screens, Text, Time series → Actions
Interaction
Agentic
Environment
Portal 2
Items from
Real work and use
Refresh
One fixed set
Run by
A host
In views
Games, Agents
Strongest model tested
GPT-6.2 Astra #1 on the capability index
Last model added
November 2026
Scores refreshed
November 20th, 2026
9
Models scored on its own leaderboard

2 models share the top score of 26.0

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.

  1. 1 GPT-6.2 Astra Max effort 26.0
  2. 2 GPT-6.2 Sol Max effort 26.0
  3. 3 GPT-6 Astra Max effort 20.0
  4. 4 GPT-6.1 Sol Max effort 20.0
  5. 5 GPT-6.1 Terra Max effort 19.0
  6. 6 Claude Opus 5.5 Max effort 11.0
  7. 7 Claude Sonnet 5.5 Max effort 6.00
  8. 8 Claude Fable 5.1 Max effort 3.00
  9. 9 GPT-6 Luna Max effort 0.00

Source: Official Portal2Bench leaderboard

#ModelScoreSettingDateSource
1 GPT-6.2 Astra 26.0 Max effort, Codex Nov 2026 Official leaderboard
2 GPT-6.2 Sol 26.0 Max effort, Codex Nov 2026 Official leaderboard
3 GPT-6 Astra 20.0 Max effort, Codex Sep 2026 Official leaderboard
4 GPT-6.1 Sol 20.0 Max effort, Codex Sep 2026 Official leaderboard
5 GPT-6.1 Terra 19.0 Max effort, Codex Oct 2026 Official leaderboard
6 Claude Opus 5.5 11.0 Max effort, Claude Code Sep 2026 Official leaderboard
7 Claude Sonnet 5.5 6.00 Max effort, Claude Code Sep 2026 Official leaderboard
8 Claude Fable 5.1 3.00 Max effort, Claude Code Sep 2026 Official leaderboard
9 GPT-6 Luna 0.00 Max effort, Codex Sep 2026 Official leaderboard

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

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

OpenAI
Own leaderboard 2 leaders

Source: Official Portal2Bench 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 10.0 20.0 30.0 2027 0.00 10.0 20.0 30.0 2027

Source: Official Portal2Bench 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

9 of 10 checks pass, 1 warning. Look first at the Task access check. We don’t know whether the tasks are public. Every benchmark

Catalog record 1 warning 1 pass
  • Warning Task access We don’t know whether the 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 9 models we track.
  • Pass Newest score The newest is from November 2026, for GPT-6.2 Astra.
  • Pass Frontier models 7 of the top 10 have a score, led by GPT-6.2 Astra (#1).
  • Note Headroom Scored as a plain number, with no fixed ceiling.
Leaderboard import 3 pass
  • Pass Importer We read the board on kingstonyip.com: 9 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