Benchmark Almanac

WordleArenas

Claude Opus 4 100% leads since February 2026
Not in model cards

Models guess a five-letter word from colored clues. Each arena changes how many clues give false feedback after a guess, from zero to five.

Liu reports 41 models across 3,749 games in the paper. The later batch results cover 52 configurations across 4,163 games with different trial counts.

Made by
Chang Liu
Released
February 2026
Kind
Puzzles and chess
Level
Everyday
Scored
Checked automatically
Answers
Agent task
Tasks
All public
Measures
Capability
Language
English
In and out
Text → Actions
Interaction
Agentic
Environment
Wordle
Items from
Generated
Refresh
One fixed set
Run by
Anyone
In views
Games, Agents
Strongest model tested
Gemini 3.1 Pro #40 on the capability index
Last model added
April 2026
Scores refreshed
November 20th, 2026
45
Entries on its own leaderboard, 32 models we track

2 models share the top score of 100% among models released by April 2026 2 models share the top score of 100% among models released by April 2026

Best score per entry from the benchmark's own leaderboard, read on November 20th, 2026. Where a model ran with several settings or harnesses, we keep its best. Every run is in the next chart. Names in gray aren't in our model list: fine-tunes, agent systems, products and older models. They don't count toward leaders. This view shows only models we track, each at its best.

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

  1. 1 Claude Opus 4 100%
  2. 2 Claude Sonnet 4 100%
  3. 3 DeepSeek-V4-Pro 95%
  4. 4 DeepSeek-R1 62.2%
  5. 5 DeepSeek-V4-Flash 57.2%
  6. 6 Gemma 4 31B IT 51.7%
  7. 7 Gemini 3 Flash 42%
  8. 8 Glm 5 cloud 41.9%
  9. 9 Kimi K2.5 cloud 39.4%
  10. 10 Gemini 3.1 Pro 38.3%
  11. 11 Qwen3 30b 30.7%
  12. 12 Qwen3.5-122B-A10B 25.3%
  13. 13 Gemini 2.5 Flash Sep 2025 25%
  14. 14 Qwen3.5 cloud 23.2%
  15. 15 Gemma 4 26B A4B 21.7%
  16. 16 Qwen3-14B 21.3%
  17. 17 Gemini 2.5 Pro Jun 2025 19.7%
  18. 18 Qwen3-8B 14.3%
  19. 19 o3 13.9%
  20. 20 Claude Haiku 4.5 13.1%
  21. 21 Qwen3-32B 12.8%
  22. 22 Qwen3.5-27B 11.7%
  23. 23 GPT-5 mini 9.80%
  24. 24 Claude Sonnet 4.6 9.50%
  25. 25 Deepseek Chat 9.20%
  26. 26 GPT-4o Mar 2025 8.80%
  27. 27 Gemma 3 27B 7.30%
  28. 28 GPT-5 7.30%
  29. 29 GPT-5.1 7.20%
  30. 30 Llama3.1 5.70%
  31. 31 Gemini 2.0 Flash Feb 2025 5.60%
  32. 32 Llama4 3.80%
  33. 33 GPT-4o mini 3.50%
  34. 34 Gemma 4 E2B 1.70%
  35. 35 Minimax M2.5 cloud 1.30%
  36. 36 Codex Mini 0%
  37. 37 Deepseek V2 16b 0%
  38. 38 Llama3 0%
  39. 39 Llama3.2 0%
  40. 40 SmolLM3 3b 0%
  41. 41 SmolLM2-1.7B 0%
  42. 42 SmolLM2-135M 0%
  43. 43 o4-mini 0%
  44. 44 Qwen3.5-35B-A3B 0%
  45. 45 Gemma 4 E4B 0%

Source: Official WordleArenas leaderboard

#EntryScoreSettingDateSource
1 Claude Opus 4 100% May 2025 Official leaderboard
2 Claude Sonnet 4 100% May 2025 Official leaderboard
3 DeepSeek-V4-Pro 95% Apr 2026 Official leaderboard
4 DeepSeek-R1 62.2% Jan 2025 Official leaderboard
5 DeepSeek-V4-Flash 57.2% Apr 2026 Official leaderboard
6 Gemma 4 31B IT 51.7% Apr 2026 Official leaderboard
7 Gemini 3 Flash 42% Dec 2025 Official leaderboard
8 Glm 5 cloud 41.9% Official leaderboard
9 Kimi K2.5 cloud 39.4% Official leaderboard
10 Gemini 3.1 Pro 38.3% Feb 2026 Official leaderboard
11 Qwen3 30b 30.7% Official leaderboard
12 Qwen3.5-122B-A10B 25.3% Feb 2026 Official leaderboard
13 Gemini 2.5 Flash Sep 2025 25% Sep 2025 Official leaderboard
14 Qwen3.5 cloud 23.2% Official leaderboard
15 Gemma 4 26B A4B 21.7% Apr 2026 Official leaderboard
16 Qwen3-14B 21.3% Apr 2025 Official leaderboard
17 Gemini 2.5 Pro Jun 2025 19.7% Jun 2025 Official leaderboard
18 Qwen3-8B 14.3% Apr 2025 Official leaderboard
19 o3 13.9% Apr 2025 Official leaderboard
20 Claude Haiku 4.5 13.1% Oct 2025 Official leaderboard
21 Qwen3-32B 12.8% Apr 2025 Official leaderboard
22 Qwen3.5-27B 11.7% Feb 2026 Official leaderboard
23 GPT-5 mini 9.80% Aug 2025 Official leaderboard
24 Claude Sonnet 4.6 9.50% Feb 2026 Official leaderboard
25 Deepseek Chat 9.20% Official leaderboard
26 GPT-4o Mar 2025 8.80% Mar 2025 Official leaderboard
27 Gemma 3 27B 7.30% Mar 2025 Official leaderboard
28 GPT-5 7.30% Aug 2025 Official leaderboard
29 GPT-5.1 7.20% Nov 2025 Official leaderboard
30 Llama3.1 5.70% Official leaderboard
31 Gemini 2.0 Flash Feb 2025 5.60% Feb 2025 Official leaderboard
32 Llama4 3.80% Official leaderboard
33 GPT-4o mini 3.50% Jul 2024 Official leaderboard
34 Gemma 4 E2B 1.70% Mar 2026 Official leaderboard
35 Minimax M2.5 cloud 1.30% Official leaderboard
36 Codex Mini 0% Official leaderboard
37 Deepseek V2 16b 0% Official leaderboard
38 Llama3 0% Official leaderboard
39 Llama3.2 0% Official leaderboard
40 SmolLM3 3b 0% Official leaderboard
41 SmolLM2-1.7B 0% Oct 2024 Official leaderboard
42 SmolLM2-135M 0% Oct 2024 Official leaderboard
43 o4-mini 0% Apr 2025 Official leaderboard
44 Qwen3.5-35B-A3B 0% Feb 2026 Official leaderboard
45 Gemma 4 E4B 0% Mar 2026 Official leaderboard
#ModelScoreSettingDateSource
1 Claude Opus 4 100% May 2025 Official leaderboard
2 Claude Sonnet 4 100% May 2025 Official leaderboard
3 DeepSeek-V4-Pro 95% Apr 2026 Official leaderboard
4 DeepSeek-R1 62.2% Jan 2025 Official leaderboard
5 DeepSeek-V4-Flash 57.2% Apr 2026 Official leaderboard
6 Gemma 4 31B IT 51.7% Apr 2026 Official leaderboard
7 Gemini 3 Flash 42% Dec 2025 Official leaderboard
8 Gemini 3.1 Pro 38.3% Feb 2026 Official leaderboard
9 Qwen3.5-122B-A10B 25.3% Feb 2026 Official leaderboard
10 Gemini 2.5 Flash Sep 2025 25% Sep 2025 Official leaderboard
11 Gemma 4 26B A4B 21.7% Apr 2026 Official leaderboard
12 Qwen3-14B 21.3% Apr 2025 Official leaderboard
13 Gemini 2.5 Pro Jun 2025 19.7% Jun 2025 Official leaderboard
14 Qwen3-8B 14.3% Apr 2025 Official leaderboard
15 o3 13.9% Apr 2025 Official leaderboard
16 Claude Haiku 4.5 13.1% Oct 2025 Official leaderboard
17 Qwen3-32B 12.8% Apr 2025 Official leaderboard
18 Qwen3.5-27B 11.7% Feb 2026 Official leaderboard
19 GPT-5 mini 9.80% Aug 2025 Official leaderboard
20 Claude Sonnet 4.6 9.50% Feb 2026 Official leaderboard
21 GPT-4o Mar 2025 8.80% Mar 2025 Official leaderboard
22 Gemma 3 27B 7.30% Mar 2025 Official leaderboard
23 GPT-5 7.30% Aug 2025 Official leaderboard
24 GPT-5.1 7.20% Nov 2025 Official leaderboard
25 Gemini 2.0 Flash Feb 2025 5.60% Feb 2025 Official leaderboard
26 GPT-4o mini 3.50% Jul 2024 Official leaderboard
27 Gemma 4 E2B 1.70% Mar 2026 Official leaderboard
28 SmolLM2-1.7B 0% Oct 2024 Official leaderboard
29 SmolLM2-135M 0% Oct 2024 Official leaderboard
30 o4-mini 0% Apr 2025 Official leaderboard
31 Qwen3.5-35B-A3B 0% Feb 2026 Official leaderboard
32 Gemma 4 E4B 0% Mar 2026 Official leaderboard

Rankings by part

We read 6 parts from the benchmark’s own leaderboard, each ranking models on one part of the test, such as Wordle, Fibble and Fibble². DeepSeek-R1 leads 4 of them, and 2 other models lead the rest. Arrows show how far a model moved from its overall rank among the models we track.

Wordle

31 models we track have a score here.

  1. 1 Claude Opus 4 100%
  2. 2 Claude Sonnet 4 100%
  3. 3 Gemini 3 Flash ↑4 100%
  4. 4 Gemma 4 26B A4B ↑7 100%
  5. 5 Gemma 4 31B IT ↑1 100%
  6. 6 DeepSeek-V4-Pro ↓3 100%
  7. 7 Gemini 3.1 Pro ↑1 95%
  8. 8 Qwen3-14B ↑4 91.7%
  9. 9 DeepSeek-R1 ↓5 90%
  10. 10 Gemini 2.5 Pro Jun 2025 ↑3 90%

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
DeepSeek-V4-Flash
no_think
57.2%
Qwen3.5-122B-A10B
no_think
25.3%
Qwen3-8B
no_think
14.3%
Qwen3-32B
no_think
12.8%
Qwen3.5-27B
no_think
11.7%
Gemma 3 27B
no_think
7.30%

Source: Official WordleArenas 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
DeepSeek-V4-Flash
no_think
57.2%
Qwen3.5-122B-A10B
no_think
25.3%
Qwen3-8B
no_think
14.3%
Qwen3-32B
no_think
12.8%
Qwen3.5-27B
no_think
11.7%
Gemma 3 27B
no_think
7.30%

Source: Official WordleArenas leaderboard

Number one over time

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

Anthropic
Own leaderboard Passed 85% in Feb 2026, by Claude Opus 4

Source: Official WordleArenas leaderboard

Claude Opus 4 on the official leaderboard passed 85% in May 2025

Models released before it came out already scored that high. 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 Best so far
0% 20% 40% 60% 80% 100% 2025 2026 Near the ceiling (85%) Claude Opus 4 DeepSeek-R1 GPT-4o mini 0% 20% 40% 60% 80% 100% 2025 2026 Near the ceiling (85%) Claude Opus 4 DeepSeek-R1

Source: Official WordleArenas 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 8 checks pass. Every benchmark

Catalog record 2 pass
  • Pass Task access All tasks are public.
  • Pass Confidence Checked against a primary source.
Scores 2 pass 3 notes
  • Pass Independent runs Scores from the official leaderboard (45 entries).
  • Pass Models scored 32 models we track, plus 13 other entries on its leaderboard.
  • Note Newest score The newest came 7 months ago, in April 2026, for DeepSeek-V4-Flash. That’s expected for a saturated test.
  • Note Frontier models None of the top 10 has a score, which is fine for a retired test. The strongest tested is Gemini 3.1 Pro (#40).
  • Note Headroom Saturated. The best is 100%, past 85% since May 2025.
Leaderboard import 3 pass
  • Pass Importer We read the board on drchangliu.github.io: 38 rows for models we track and 13 other entries 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