Benchmark Almanac

PokéBench: Tool Assisted Benchmark

GPT-5.5 10.0 leads since September 2026
Not in model cards

Models play Pokémon Red with a screenshot, map, game text, and action shortcuts. They have 1,000 turns to reach ten milestones ending with Brock.

The maker ranks 41 runs under 40 display names. Two provider IDs share the DeepSeek V4.1 Flash label. The runs span two prompt versions.

Made by
Cody Harris
Released
September 2026
Kind
Video games
Level
Everyday
Scored
Checked automatically
Answers
Agent task
Tasks
All public
Measures
Capability
Language
English
In and out
Screens, Text → Actions
Interaction
Agentic
Environment
Pokémon Red and Blue
Items from
Real work and use
Run by
Anyone
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
38
Entries on its own leaderboard, 34 models we track

20 models share the top score of 10.0 20 models share the top score of 10.0

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.

  1. 1 GPT-5.5 gpt-5.5, prompt v22, high 10.0
  2. 2 GPT-5.6 Sol openai/gpt-5.6-sol, prompt v21, high 10.0
  3. 3 Grok 4.6 grok-4.6, prompt v21, high 10.0
  4. 4 Claude Fable 5.1 claude-fable-5-1, prompt v22, high 10.0
  5. 5 Gemini 3.8 Flash gemini-3.8-flash, prompt v21, high 10.0
  6. 6 GPT-6 Astra gpt-6-astra-1, prompt v21, high 10.0
  7. 7 Grok 4.7 grok-4.7, prompt v21, high 10.0
  8. 8 Claude Opus 5.5 claude-opus-5-5, prompt v22, high 10.0
  9. 9 GPT-6 Sol gpt-6-sol, prompt v22, high 10.0
  10. 10 Claude Sonnet 5.5 claude-sonnet-5-5, prompt v22, high 10.0
  11. 11 GPT-6.1 Sol gpt-6.1-sol, prompt v22, high 10.0
  12. 12 Gemini 4 Argon gemini-4-argon, prompt v22, high 10.0
  13. 13 GPT-6.1 Terra gpt-6.1-terra, prompt v22, high 10.0
  14. 14 Gemini 4 Carbon gemini-4-carbon, prompt v22, high 10.0
  15. 15 Tencent Hy4 tencent/hy4, prompt v22, high, fp4 10.0
  16. 16 Muse Ember muse-ember-1.0, prompt v22, high 10.0
  17. 17 Gemini 4 Hydrogen gemini-4-hydrogen, prompt v22, high 10.0
  18. 18 Qwen4-Max qwen4-max-preview, prompt v22, high 10.0
  19. 19 GPT-6.2 Astra gpt-6.2-astra, prompt v22, high 10.0
  20. 20 Hy4 Flash hy4-flash-dev, prompt v22, high 10.0
  21. 21 Gemini 4 Neon gemini-4-neon, prompt v22, high 10.0
  22. 22 Claude Haiku 5.5 claude-haiku-5-5, prompt v22, high 9.00
  23. 23 DeepSeek V4.1 Flash deepseek-flash, prompt v21, high 8.00
  24. 24 GPT-6 Luna gpt-6-luna, prompt v22, high 8.00
  25. 25 GPT-5.6 Luna openai/gpt-5.6-luna, prompt v21, high 7.00
  26. 26 GLM-5.3-Flash z-ai/glm-5.3-flash-20260826, prompt v21, high 7.00
  27. 27 Muse Spark 1.3 meta/muse-spark-1.3-contributor, prompt v21, high 7.00
  28. 28 Muse Wonder meta/muse-wonder, prompt v22, high 7.00
  29. 29 Crouton Alpha stealth/crouton-alpha, prompt v22, high 7.00
  30. 30 Grok 4.1 Fast grok-4-1-fast-reasoning, prompt v21, high 6.00
  31. 31 Grok 4.20 grok-4-20-reasoning, prompt v21, high 4.00
  32. 32 Gemma 4 31B IT gemma4:31b, prompt v22, high, Q4_K_M 4.00
  33. 33 Grok 4.3 grok-4.3, prompt v21, high 4.00
  34. 34 Qwen3.8 Flash qwen/qwen3.8-flash-20260826, prompt v22, high 4.00
  35. 35 Space Bunny Alpha stealth/space-bunny-alpha, prompt v22, high 4.00
  36. 36 Muse Glimmer muse-glimmer:30b, prompt v21, high, Q4_K_M 3.00
  37. 37 Qwen 3.8 27B qwen3.8:27b, prompt v22, high, Q4_K_M 1.00
  38. 38 Qwen3-VL 30B qwen3-vl:30b, prompt v21, high, Q4_K_M 1.00

Source: Official PokéBench: Tool Assisted Benchmark leaderboard

  1. 1 GPT-5.5 gpt-5.5, prompt v22, high 10.0
  2. 2 GPT-5.6 Sol openai/gpt-5.6-sol, prompt v21, high 10.0
  3. 3 Grok 4.6 grok-4.6, prompt v21, high 10.0
  4. 4 Claude Fable 5.1 claude-fable-5-1, prompt v22, high 10.0
  5. 5 Gemini 3.8 Flash gemini-3.8-flash, prompt v21, high 10.0
  6. 6 GPT-6 Astra gpt-6-astra-1, prompt v21, high 10.0
  7. 7 Grok 4.7 grok-4.7, prompt v21, high 10.0
  8. 8 Claude Opus 5.5 claude-opus-5-5, prompt v22, high 10.0
  9. 9 GPT-6 Sol gpt-6-sol, prompt v22, high 10.0
  10. 10 Claude Sonnet 5.5 claude-sonnet-5-5, prompt v22, high 10.0
  11. 11 GPT-6.1 Sol gpt-6.1-sol, prompt v22, high 10.0
  12. 12 Gemini 4 Argon gemini-4-argon, prompt v22, high 10.0
  13. 13 GPT-6.1 Terra gpt-6.1-terra, prompt v22, high 10.0
  14. 14 Gemini 4 Carbon gemini-4-carbon, prompt v22, high 10.0
  15. 15 Tencent Hy4 tencent/hy4, prompt v22, high, fp4 10.0
  16. 16 Muse Ember muse-ember-1.0, prompt v22, high 10.0
  17. 17 Gemini 4 Hydrogen gemini-4-hydrogen, prompt v22, high 10.0
  18. 18 Qwen4-Max qwen4-max-preview, prompt v22, high 10.0
  19. 19 GPT-6.2 Astra gpt-6.2-astra, prompt v22, high 10.0
  20. 20 Gemini 4 Neon gemini-4-neon, prompt v22, high 10.0
  21. 21 Claude Haiku 5.5 claude-haiku-5-5, prompt v22, high 9.00
  22. 22 DeepSeek V4.1 Flash deepseek-flash, prompt v21, high 8.00
  23. 23 GPT-6 Luna gpt-6-luna, prompt v22, high 8.00
  24. 24 GPT-5.6 Luna openai/gpt-5.6-luna, prompt v21, high 7.00
  25. 25 GLM-5.3-Flash z-ai/glm-5.3-flash-20260826, prompt v21, high 7.00
  26. 26 Muse Spark 1.3 meta/muse-spark-1.3-contributor, prompt v21, high 7.00
  27. 27 Muse Wonder meta/muse-wonder, prompt v22, high 7.00
  28. 28 Grok 4.1 Fast grok-4-1-fast-reasoning, prompt v21, high 6.00
  29. 29 Grok 4.20 grok-4-20-reasoning, prompt v21, high 4.00
  30. 30 Gemma 4 31B IT gemma4:31b, prompt v22, high, Q4_K_M 4.00
  31. 31 Grok 4.3 grok-4.3, prompt v21, high 4.00
  32. 32 Qwen3.8 Flash qwen/qwen3.8-flash-20260826, prompt v22, high 4.00
  33. 33 Muse Glimmer muse-glimmer:30b, prompt v21, high, Q4_K_M 3.00
  34. 34 Qwen 3.8 27B qwen3.8:27b, prompt v22, high, Q4_K_M 1.00

Source: Official PokéBench: Tool Assisted Benchmark leaderboard

#EntryScoreSettingReleasedRun onLinkSource
1 GPT-5.5 10.0 gpt-5.5, prompt v22, high, TAB v2 Apr 2026 Oct 2026 youtube.com Official leaderboard
2 GPT-5.6 Sol 10.0 openai/gpt-5.6-sol, prompt v21, high, TAB v2 Jul 2026 Sep 2026 youtube.com Official leaderboard
3 Grok 4.6 10.0 grok-4.6, prompt v21, high, TAB v2 Aug 2026 Sep 2026 youtube.com Official leaderboard
4 Claude Fable 5.1 10.0 claude-fable-5-1, prompt v22, high, TAB v2 Sep 2026 Sep 2026 youtube.com Official leaderboard
5 Gemini 3.8 Flash 10.0 gemini-3.8-flash, prompt v21, high, TAB v2 Sep 2026 Sep 2026 youtu.be Official leaderboard
6 GPT-6 Astra 10.0 gpt-6-astra-1, prompt v21, high, TAB v2 Sep 2026 Sep 2026 youtube.com Official leaderboard
7 Grok 4.7 10.0 grok-4.7, prompt v21, high, TAB v2 Sep 2026 Sep 2026 youtube.com Official leaderboard
8 Claude Opus 5.5 10.0 claude-opus-5-5, prompt v22, high, TAB v2 Sep 2026 Sep 2026 youtube.com Official leaderboard
9 GPT-6 Sol 10.0 gpt-6-sol, prompt v22, high, TAB v2 Sep 2026 Sep 2026 youtube.com Official leaderboard
10 Claude Sonnet 5.5 10.0 claude-sonnet-5-5, prompt v22, high, TAB v2 Sep 2026 Sep 2026 youtube.com Official leaderboard
11 GPT-6.1 Sol 10.0 gpt-6.1-sol, prompt v22, high, TAB v2 Sep 2026 Sep 2026 youtube.com Official leaderboard
12 Gemini 4 Argon 10.0 gemini-4-argon, prompt v22, high, TAB v2 Sep 2026 Oct 2026 youtube.com Official leaderboard
13 GPT-6.1 Terra 10.0 gpt-6.1-terra, prompt v22, high, TAB v2 Oct 2026 Oct 2026 Official leaderboard
14 Gemini 4 Carbon 10.0 gemini-4-carbon, prompt v22, high, TAB v2 Oct 2026 Nov 2026 Official leaderboard
15 Tencent Hy4 10.0 tencent/hy4, prompt v22, high, fp4, TAB v2 Nov 2026 Nov 2026 youtube.com Official leaderboard
16 Muse Ember 10.0 muse-ember-1.0, prompt v22, high, TAB v2 Nov 2026 Nov 2026 youtube.com Official leaderboard
17 Gemini 4 Hydrogen 10.0 gemini-4-hydrogen, prompt v22, high, TAB v2 Nov 2026 Nov 2026 youtube.com Official leaderboard
18 Qwen4-Max 10.0 qwen4-max-preview, prompt v22, high, TAB v2 Nov 2026 Nov 2026 youtube.com Official leaderboard
19 GPT-6.2 Astra 10.0 gpt-6.2-astra, prompt v22, high, TAB v2 Nov 2026 Nov 2026 youtube.com Official leaderboard
20 Hy4 Flash 10.0 hy4-flash-dev, prompt v22, high, TAB v2 Nov 2026 youtube.com Official leaderboard
21 Gemini 4 Neon 10.0 gemini-4-neon, prompt v22, high, TAB v2 Nov 2026 Nov 2026 youtube.com Official leaderboard
22 Claude Haiku 5.5 9.00 claude-haiku-5-5, prompt v22, high, TAB v2 Oct 2026 Oct 2026 youtube.com Official leaderboard
23 DeepSeek V4.1 Flash 8.00 deepseek-flash, prompt v21, high, TAB v2 Sep 2026 Sep 2026 Official leaderboard
24 GPT-6 Luna 8.00 gpt-6-luna, prompt v22, high, TAB v2 Sep 2026 Sep 2026 youtube.com Official leaderboard
25 GPT-5.6 Luna 7.00 openai/gpt-5.6-luna, prompt v21, high, TAB v2 Jul 2026 Sep 2026 youtube.com Official leaderboard
26 GLM-5.3-Flash 7.00 z-ai/glm-5.3-flash-20260826, prompt v21, high, TAB v2 Aug 2026 Sep 2026 youtube.com Official leaderboard
27 Muse Spark 1.3 7.00 meta/muse-spark-1.3-contributor, prompt v21, high, TAB v2 Sep 2026 Sep 2026 youtu.be Official leaderboard
28 Muse Wonder 7.00 meta/muse-wonder, prompt v22, high, TAB v2 Oct 2026 Oct 2026 youtube.com Official leaderboard
29 Crouton Alpha 7.00 stealth/crouton-alpha, prompt v22, high, TAB v2 Nov 2026 youtube.com Official leaderboard
30 Grok 4.1 Fast 6.00 grok-4-1-fast-reasoning, prompt v21, high, TAB v2 Nov 2025 Sep 2026 youtube.com Official leaderboard
31 Grok 4.20 4.00 grok-4-20-reasoning, prompt v21, high, TAB v2 Feb 2026 Sep 2026 youtube.com Official leaderboard
32 Gemma 4 31B IT 4.00 gemma4:31b, prompt v22, high, Q4_K_M, TAB v2 Apr 2026 Sep 2026 youtube.com Official leaderboard
33 Grok 4.3 4.00 grok-4.3, prompt v21, high, TAB v2 Apr 2026 Sep 2026 youtube.com Official leaderboard
34 Qwen3.8 Flash 4.00 qwen/qwen3.8-flash-20260826, prompt v22, high, TAB v2 Aug 2026 Sep 2026 youtube.com Official leaderboard
35 Space Bunny Alpha 4.00 stealth/space-bunny-alpha, prompt v22, high, TAB v2 Sep 2026 youtube.com Official leaderboard
36 Muse Glimmer 3.00 muse-glimmer:30b, prompt v21, high, Q4_K_M, TAB v2 Aug 2026 Sep 2026 youtu.be Official leaderboard
37 Qwen 3.8 27B 1.00 qwen3.8:27b, prompt v22, high, Q4_K_M, TAB v2 Aug 2026 Sep 2026 youtube.com Official leaderboard
38 Qwen3-VL 30B 31.1B 1.00 qwen3-vl:30b, prompt v21, high, Q4_K_M, TAB v2 Sep 2026 youtube.com Official leaderboard
#ModelScoreSettingReleasedRun onLinkSource
1 GPT-5.5 10.0 gpt-5.5, prompt v22, high, TAB v2 Apr 2026 Oct 2026 youtube.com Official leaderboard
2 GPT-5.6 Sol 10.0 openai/gpt-5.6-sol, prompt v21, high, TAB v2 Jul 2026 Sep 2026 youtube.com Official leaderboard
3 Grok 4.6 10.0 grok-4.6, prompt v21, high, TAB v2 Aug 2026 Sep 2026 youtube.com Official leaderboard
4 Claude Fable 5.1 10.0 claude-fable-5-1, prompt v22, high, TAB v2 Sep 2026 Sep 2026 youtube.com Official leaderboard
5 Gemini 3.8 Flash 10.0 gemini-3.8-flash, prompt v21, high, TAB v2 Sep 2026 Sep 2026 youtu.be Official leaderboard
6 GPT-6 Astra 10.0 gpt-6-astra-1, prompt v21, high, TAB v2 Sep 2026 Sep 2026 youtube.com Official leaderboard
7 Grok 4.7 10.0 grok-4.7, prompt v21, high, TAB v2 Sep 2026 Sep 2026 youtube.com Official leaderboard
8 Claude Opus 5.5 10.0 claude-opus-5-5, prompt v22, high, TAB v2 Sep 2026 Sep 2026 youtube.com Official leaderboard
9 GPT-6 Sol 10.0 gpt-6-sol, prompt v22, high, TAB v2 Sep 2026 Sep 2026 youtube.com Official leaderboard
10 Claude Sonnet 5.5 10.0 claude-sonnet-5-5, prompt v22, high, TAB v2 Sep 2026 Sep 2026 youtube.com Official leaderboard
11 GPT-6.1 Sol 10.0 gpt-6.1-sol, prompt v22, high, TAB v2 Sep 2026 Sep 2026 youtube.com Official leaderboard
12 Gemini 4 Argon 10.0 gemini-4-argon, prompt v22, high, TAB v2 Sep 2026 Oct 2026 youtube.com Official leaderboard
13 GPT-6.1 Terra 10.0 gpt-6.1-terra, prompt v22, high, TAB v2 Oct 2026 Oct 2026 Official leaderboard
14 Gemini 4 Carbon 10.0 gemini-4-carbon, prompt v22, high, TAB v2 Oct 2026 Nov 2026 Official leaderboard
15 Tencent Hy4 10.0 tencent/hy4, prompt v22, high, fp4, TAB v2 Nov 2026 Nov 2026 youtube.com Official leaderboard
16 Muse Ember 10.0 muse-ember-1.0, prompt v22, high, TAB v2 Nov 2026 Nov 2026 youtube.com Official leaderboard
17 Gemini 4 Hydrogen 10.0 gemini-4-hydrogen, prompt v22, high, TAB v2 Nov 2026 Nov 2026 youtube.com Official leaderboard
18 Qwen4-Max 10.0 qwen4-max-preview, prompt v22, high, TAB v2 Nov 2026 Nov 2026 youtube.com Official leaderboard
19 GPT-6.2 Astra 10.0 gpt-6.2-astra, prompt v22, high, TAB v2 Nov 2026 Nov 2026 youtube.com Official leaderboard
20 Gemini 4 Neon 10.0 gemini-4-neon, prompt v22, high, TAB v2 Nov 2026 Nov 2026 youtube.com Official leaderboard
21 Claude Haiku 5.5 9.00 claude-haiku-5-5, prompt v22, high, TAB v2 Oct 2026 Oct 2026 youtube.com Official leaderboard
22 DeepSeek V4.1 Flash 8.00 deepseek-flash, prompt v21, high, TAB v2 Sep 2026 Sep 2026 Official leaderboard
23 GPT-6 Luna 8.00 gpt-6-luna, prompt v22, high, TAB v2 Sep 2026 Sep 2026 youtube.com Official leaderboard
24 GPT-5.6 Luna 7.00 openai/gpt-5.6-luna, prompt v21, high, TAB v2 Jul 2026 Sep 2026 youtube.com Official leaderboard
25 GLM-5.3-Flash 7.00 z-ai/glm-5.3-flash-20260826, prompt v21, high, TAB v2 Aug 2026 Sep 2026 youtube.com Official leaderboard
26 Muse Spark 1.3 7.00 meta/muse-spark-1.3-contributor, prompt v21, high, TAB v2 Sep 2026 Sep 2026 youtu.be Official leaderboard
27 Muse Wonder 7.00 meta/muse-wonder, prompt v22, high, TAB v2 Oct 2026 Oct 2026 youtube.com Official leaderboard
28 Grok 4.1 Fast 6.00 grok-4-1-fast-reasoning, prompt v21, high, TAB v2 Nov 2025 Sep 2026 youtube.com Official leaderboard
29 Grok 4.20 4.00 grok-4-20-reasoning, prompt v21, high, TAB v2 Feb 2026 Sep 2026 youtube.com Official leaderboard
30 Gemma 4 31B IT 4.00 gemma4:31b, prompt v22, high, Q4_K_M, TAB v2 Apr 2026 Sep 2026 youtube.com Official leaderboard
31 Grok 4.3 4.00 grok-4.3, prompt v21, high, TAB v2 Apr 2026 Sep 2026 youtube.com Official leaderboard
32 Qwen3.8 Flash 4.00 qwen/qwen3.8-flash-20260826, prompt v22, high, TAB v2 Aug 2026 Sep 2026 youtube.com Official leaderboard
33 Muse Glimmer 3.00 muse-glimmer:30b, prompt v21, high, Q4_K_M, TAB v2 Aug 2026 Sep 2026 youtu.be Official leaderboard
34 Qwen 3.8 27B 1.00 qwen3.8:27b, prompt v22, high, Q4_K_M, TAB v2 Aug 2026 Sep 2026 youtube.com Official leaderboard

The best score, 10.0, costs $1.14 for the full run

Runs on the cost frontier score higher than any run that's cheaper. The costs are the leaderboard's own estimates, so compare them only within this benchmark.

Meta Cost frontier One model at each effort level
0.00 2.00 4.00 6.00 8.00 10.0 $1.00 $2.00 $5.00 $10 $20 $50 $100 Cost of the full run, log scale Muse Ember, muse-ember-1.0, prompt v22, high 0.00 2.00 4.00 6.00 8.00 10.0 $1.00 $10 $100 Cost of the full run, log scale Muse Ember, muse-ember-1.0, prompt v22, high

Source: Official PokéBench: Tool Assisted Benchmark leaderboard

ModelEffortHarnessScoreCost of the full runOutput tokens
GPT-5.5 gpt-5.5, prompt v22, high TAB v2 10.0 $47 740,247
GPT-5.6 Sol openai/gpt-5.6-sol, prompt v21, high TAB v2 10.0 $9.66 139,459
Grok 4.6 grok-4.6, prompt v21, high TAB v2 10.0 $10 1,081,083
Claude Fable 5.1 claude-fable-5-1, prompt v22, high TAB v2 10.0 $100 354,706
Gemini 3.8 Flash gemini-3.8-flash, prompt v21, high TAB v2 10.0 $3.70 627,490
GPT-6 Astra gpt-6-astra-1, prompt v21, high TAB v2 10.0 $13 41,588
Grok 4.7 grok-4.7, prompt v21, high TAB v2 10.0 $22 1,804,385
Claude Opus 5.5 claude-opus-5-5, prompt v22, high TAB v2 10.0 $8.44 114,600
GPT-6 Sol gpt-6-sol, prompt v22, high TAB v2 10.0 $16 226,480
Claude Sonnet 5.5 claude-sonnet-5-5, prompt v22, high TAB v2 10.0 $10 210,905
GPT-6.1 Sol gpt-6.1-sol, prompt v22, high TAB v2 10.0 $2.60 47,752
Gemini 4 Argon gemini-4-argon, prompt v22, high TAB v2 10.0 $3.20 97,922
GPT-6.1 Terra gpt-6.1-terra, prompt v22, high TAB v2 10.0 $5.47 63,249
Gemini 4 Carbon gemini-4-carbon, prompt v22, high TAB v2 10.0 $6.70 394,211
Tencent Hy4 tencent/hy4, prompt v22, high, fp4 TAB v2 10.0 $12 8,512,355
Muse Ember muse-ember-1.0, prompt v22, high TAB v2 10.0 $1.14 652,674
Gemini 4 Hydrogen gemini-4-hydrogen, prompt v22, high TAB v2 10.0 278,299
Qwen4-Max qwen4-max-preview, prompt v22, high TAB v2 10.0 1,697,270
GPT-6.2 Astra gpt-6.2-astra, prompt v22, high TAB v2 10.0 67,493
Gemini 4 Neon gemini-4-neon, prompt v22, high TAB v2 10.0 $7.06 149,033
Claude Haiku 5.5 claude-haiku-5-5, prompt v22, high TAB v2 9.00 $2.36 1,418,648
DeepSeek V4.1 Flash deepseek-flash, prompt v21, high TAB v2 8.00 $13 6,564,832
GPT-6 Luna gpt-6-luna, prompt v22, high TAB v2 8.00 $1.85 557,004
GPT-5.6 Luna openai/gpt-5.6-luna, prompt v21, high TAB v2 7.00 $3.04 596,496
GLM-5.3-Flash z-ai/glm-5.3-flash-20260826, prompt v21, high TAB v2 7.00 $2.48 3,742,751
Muse Spark 1.3 meta/muse-spark-1.3-contributor, prompt v21, high TAB v2 7.00 $2.13 2,134,047
DeepSeek V4.1 Flash deepseek/deepseek-v4.1-flash, prompt v22, high TAB v2 7.00 $6.70 6,272,745
Muse Wonder meta/muse-wonder, prompt v22, high TAB v2 7.00 $96 1,244,424
Grok 4.1 Fast grok-4-1-fast-reasoning, prompt v21, high TAB v2 6.00 $4.39 4,160,995
Grok 4.20 grok-4-20-reasoning, prompt v21, high TAB v2 4.00 $31 5,202,069
Gemma 4 31B IT gemma4:31b, prompt v22, high, Q4_K_M TAB v2 4.00 89,201
Grok 4.3 grok-4.3, prompt v21, high TAB v2 4.00 $37 2,504,778
Qwen3.8 Flash qwen/qwen3.8-flash-20260826, prompt v22, high TAB v2 4.00 $5.95 5,335,815
Grok 4.1 Fast grok-4-1-fast-non-reasoning, prompt v21, default TAB v2 3.00 $4.06 139,170
Muse Glimmer muse-glimmer:30b, prompt v21, high, Q4_K_M TAB v2 3.00 137,181
Grok 4.20 grok-4-20-non-reasoning, prompt v21, default TAB v2 2.00 $30 211,330
Qwen 3.8 27B qwen3.8:27b, prompt v22, high, Q4_K_M TAB v2 1.00 174,653

The best score, 10.0, costs $1.14 for the full run

Runs on the cost frontier score higher than any run that's cheaper. The costs are the leaderboard's own estimates, so compare them only within this benchmark.

Meta Cost frontier One model at each effort level
0.00 2.00 4.00 6.00 8.00 10.0 $1.00 $2.00 $5.00 $10 $20 $50 $100 Cost of the full run, log scale Muse Ember, muse-ember-1.0, prompt v22, high 0.00 2.00 4.00 6.00 8.00 10.0 $1.00 $10 $100 Cost of the full run, log scale Muse Ember, muse-ember-1.0, prompt v22, high

Source: Official PokéBench: Tool Assisted Benchmark leaderboard

ModelEffortHarnessScoreCost of the full runOutput tokens
GPT-5.5 gpt-5.5, prompt v22, high TAB v2 10.0 $47 740,247
GPT-5.6 Sol openai/gpt-5.6-sol, prompt v21, high TAB v2 10.0 $9.66 139,459
Grok 4.6 grok-4.6, prompt v21, high TAB v2 10.0 $10 1,081,083
Claude Fable 5.1 claude-fable-5-1, prompt v22, high TAB v2 10.0 $100 354,706
Gemini 3.8 Flash gemini-3.8-flash, prompt v21, high TAB v2 10.0 $3.70 627,490
GPT-6 Astra gpt-6-astra-1, prompt v21, high TAB v2 10.0 $13 41,588
Grok 4.7 grok-4.7, prompt v21, high TAB v2 10.0 $22 1,804,385
Claude Opus 5.5 claude-opus-5-5, prompt v22, high TAB v2 10.0 $8.44 114,600
GPT-6 Sol gpt-6-sol, prompt v22, high TAB v2 10.0 $16 226,480
Claude Sonnet 5.5 claude-sonnet-5-5, prompt v22, high TAB v2 10.0 $10 210,905
GPT-6.1 Sol gpt-6.1-sol, prompt v22, high TAB v2 10.0 $2.60 47,752
Gemini 4 Argon gemini-4-argon, prompt v22, high TAB v2 10.0 $3.20 97,922
GPT-6.1 Terra gpt-6.1-terra, prompt v22, high TAB v2 10.0 $5.47 63,249
Gemini 4 Carbon gemini-4-carbon, prompt v22, high TAB v2 10.0 $6.70 394,211
Tencent Hy4 tencent/hy4, prompt v22, high, fp4 TAB v2 10.0 $12 8,512,355
Muse Ember muse-ember-1.0, prompt v22, high TAB v2 10.0 $1.14 652,674
Gemini 4 Hydrogen gemini-4-hydrogen, prompt v22, high TAB v2 10.0 278,299
Qwen4-Max qwen4-max-preview, prompt v22, high TAB v2 10.0 1,697,270
GPT-6.2 Astra gpt-6.2-astra, prompt v22, high TAB v2 10.0 67,493
Gemini 4 Neon gemini-4-neon, prompt v22, high TAB v2 10.0 $7.06 149,033
Claude Haiku 5.5 claude-haiku-5-5, prompt v22, high TAB v2 9.00 $2.36 1,418,648
DeepSeek V4.1 Flash deepseek-flash, prompt v21, high TAB v2 8.00 $13 6,564,832
GPT-6 Luna gpt-6-luna, prompt v22, high TAB v2 8.00 $1.85 557,004
GPT-5.6 Luna openai/gpt-5.6-luna, prompt v21, high TAB v2 7.00 $3.04 596,496
GLM-5.3-Flash z-ai/glm-5.3-flash-20260826, prompt v21, high TAB v2 7.00 $2.48 3,742,751
Muse Spark 1.3 meta/muse-spark-1.3-contributor, prompt v21, high TAB v2 7.00 $2.13 2,134,047
DeepSeek V4.1 Flash deepseek/deepseek-v4.1-flash, prompt v22, high TAB v2 7.00 $6.70 6,272,745
Muse Wonder meta/muse-wonder, prompt v22, high TAB v2 7.00 $96 1,244,424
Grok 4.1 Fast grok-4-1-fast-reasoning, prompt v21, high TAB v2 6.00 $4.39 4,160,995
Grok 4.20 grok-4-20-reasoning, prompt v21, high TAB v2 4.00 $31 5,202,069
Gemma 4 31B IT gemma4:31b, prompt v22, high, Q4_K_M TAB v2 4.00 89,201
Grok 4.3 grok-4.3, prompt v21, high TAB v2 4.00 $37 2,504,778
Qwen3.8 Flash qwen/qwen3.8-flash-20260826, prompt v22, high TAB v2 4.00 $5.95 5,335,815
Grok 4.1 Fast grok-4-1-fast-non-reasoning, prompt v21, default TAB v2 3.00 $4.06 139,170
Muse Glimmer muse-glimmer:30b, prompt v21, high, Q4_K_M TAB v2 3.00 137,181
Grok 4.20 grok-4-20-non-reasoning, prompt v21, default TAB v2 2.00 $30 211,330
Qwen 3.8 27B qwen3.8:27b, prompt v22, high, Q4_K_M TAB v2 1.00 174,653

Number one over time

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

OpenAI
Own leaderboard 1 leader
GPT-5.5 10.0

Source: Official PokéBench: Tool Assisted Benchmark 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 2.00 4.00 6.00 8.00 10.0 2026 2027 0.00 2.00 4.00 6.00 8.00 10.0 2026 2027

Source: Official PokéBench: Tool Assisted Benchmark 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 (38 entries).
  • Pass Models scored 34 models we track, plus 4 other entries on its leaderboard.
  • Pass Newest score The newest is from November 2026, for Gemini 4 Neon.
  • Pass Frontier models 9 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 pokebench.tv: 37 rows for models we track and 4 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