Benchmark Almanac

Android Bench 2.0

Claude Sonnet 5.5 40% leads since September 2026
Not in model cards

Coding agents complete 30 long Android tasks such as building an app from mocks, library migrations and Flutter to native conversion. Tests and an LLM visual judge grade each run.

It reports pass rate plus a continuous completion rate, and runs models in their own harnesses such as Claude Code and Codex. The task set is private to limit contamination.

Released
September 2026
Kind
Building whole projects
Level
Expert
Scored
Checked automatically
Answers
Agent task
Tasks
Kept private
Strongest model tested
Gemini 4 Carbon #2 on the capability index
Last model added
October 2026
Scores refreshed
November 12th, 2026
17
Models scored on its own leaderboard

Claude Sonnet 5.5 leads with 40%

Best score per model from the benchmark's own leaderboard, read on November 12th, 2026. Where a model ran with several settings or harnesses, we keep its best. Every run is in the next chart.

17 models. Scroll sideways for the rest.

  1. 40.0 Claude Sonnet 5.5
  2. 37.3 Claude Opus 5.5
  3. 29.3 GPT-6.1 Sol
  4. 28.7 GPT-6.1 Terra
  5. 28.0 GPT-6 Astra
  6. 24.0 Kimi K3.1 Pro
  7. 23.3 Gemini 4 Carbon
  8. 22.7 Claude Fable 5.1
  9. 21.3 Gemini 4 Argon
  10. 20.7 Claude Haiku 5.5
  11. 17.3 DeepSeek V4.1 Flash 1017
  12. 13.3 Muse Wonder
  13. 12.5 Qwen 3.8 Max
  14. 12.0 Kimi K3
  15. 10.7 GPT-6 Luna
  16. 8.00 Gemini 3.8 Flash
  17. 7.30 Gemini 3.7 Flash
  1. 1 Claude Sonnet 5.5 max 40%
  2. 2 Claude Opus 5.5 max 37.3%
  3. 3 GPT-6.1 Sol max 29.3%
  4. 4 GPT-6.1 Terra max 28.7%
  5. 5 GPT-6 Astra high 28%
  6. 6 Kimi K3.1 Pro max 24%
  7. 7 Gemini 4 Carbon high 23.3%
  8. 8 Claude Fable 5.1 high 22.7%
  9. 9 Gemini 4 Argon high 21.3%
  10. 10 Claude Haiku 5.5 max 20.7%
  11. 11 DeepSeek V4.1 Flash 1017 high 17.3%
  12. 12 Muse Wonder max 13.3%
  13. 13 Qwen 3.8 Max 12.5%
  14. 14 Kimi K3 max 12%
  15. 15 GPT-6 Luna 10.7%
  16. 16 Gemini 3.8 Flash 8.00%
  17. 17 Gemini 3.7 Flash 7.30%

Source: Official Android Bench 2.0 leaderboard

#ModelScoreSettingDateSource
1 Claude Sonnet 5.5 40% max, claude-code Sep 2026 Official leaderboard
2 Claude Opus 5.5 37.3% max, claude-code Sep 2026 Official leaderboard
3 GPT-6.1 Sol 29.3% max, codex Sep 2026 Official leaderboard
4 GPT-6.1 Terra 28.7% max, codex Oct 2026 Official leaderboard
5 GPT-6 Astra 28% high, codex Sep 2026 Official leaderboard
6 Kimi K3.1 Pro 24% max, kimi-code Oct 2026 Official leaderboard
7 Gemini 4 Carbon 23.3% high, antigravity-sdk Oct 2026 Official leaderboard
8 Claude Fable 5.1 22.7% high, claude-code Sep 2026 Official leaderboard
9 Gemini 4 Argon 21.3% high, antigravity-sdk Sep 2026 Official leaderboard
10 Claude Haiku 5.5 20.7% max, claude-code Oct 2026 Official leaderboard
11 DeepSeek V4.1 Flash 1017 17.3% high, opencode Oct 2026 Official leaderboard
12 Muse Wonder 13.3% max, muse-code Oct 2026 Official leaderboard
13 Qwen 3.8 Max 12.5% qwen-coder Aug 2026 Official leaderboard
14 Kimi K3 12% max, kimi-code Jul 2026 Official leaderboard
15 GPT-6 Luna 10.7% codex Sep 2026 Official leaderboard
16 Gemini 3.8 Flash 8.00% antigravity-sdk Sep 2026 Official leaderboard
17 Gemini 3.7 Flash 7.30% antigravity-sdk Aug 2026 Official leaderboard

The best score, 40%, costs $1,575 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.

Anthropic OpenAI DeepSeek Cost frontier One model at each effort level
0% 10% 20% 30% 40% $1.00 $10 $100 $1,000 $10,000 Cost of the full run, log scale Claude Sonnet 5.5, max Claude Opus 5.5, max GPT-6.1 Sol, max DeepSeek V4.1 Flash 1017, high GPT-6 Luna 0% 10% 20% 30% 40% $1.00 $10 $100 $1,000 $10,000 Cost of the full run, log scale Claude Sonnet 5.5, max Claude Opus 5.5, max

Source: Official Android Bench 2.0 leaderboard

ModelEffortHarnessScoreCost of the full runOutput tokens
Claude Sonnet 5.5 max claude-code 40% $1,575
Claude Opus 5.5 max claude-code 37.3% $1,150
GPT-6.1 Sol max codex 29.3% $100
GPT-6.1 Terra max codex 28.7% $178
GPT-6 Astra high codex 28% $376
Kimi K3.1 Pro max kimi-code 24% $317
Gemini 4 Carbon high antigravity-sdk 23.3% $184
Claude Fable 5.1 high claude-code 22.7% $493
Gemini 4 Argon high antigravity-sdk 21.3% $268
Claude Haiku 5.5 max claude-code 20.7% $191
DeepSeek V4.1 Flash 1017 high opencode 17.3% $9.90
Muse Wonder max muse-code 13.3% $272
Qwen 3.8 Max qwen-coder 12.5% $258
Kimi K3 max kimi-code 12% $418
GPT-6 Luna codex 10.7% $5.30
Gemini 3.8 Flash antigravity-sdk 8.00% $35
Gemini 3.7 Flash antigravity-sdk 7.30% $26

4 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 4 leaders
Claude Sonnet 5.5 40%

Source: Official Android Bench 2.0 leaderboard

Scores over time

All the scores we have, one dot each, so a model tested at several settings appears more than once.

On the benchmark’s own leaderboard Best so far
0% 20% 40% 60% 80% 100% 2027 Near the ceiling (85%) Claude Sonnet 5.5 GPT-6 Astra Claude Fable 5.1 Qwen 3.8 Max 0% 20% 40% 60% 80% 100% 2027 Near the ceiling (85%) Claude Sonnet 5.5 GPT-6 Astra

Source: Official Android Bench 2.0 leaderboard

No lab model card reports it yet

We read the headline results of 522 launch posts from 52 labs, and none of them list this benchmark.

Versions

We switch the top row to each new version when it comes out. Open it for one row per version. An old version stays in play while people still test it, so whoever first takes it past 85% gets credit for closing it out.

Google DeepMind OpenAI Anthropic
Android Bench 2.0 8 leaders
Claude Sonnet 5.5 40%
AndroidBench Passed 85% in Jun 2026, by Claude Fable 5
Claude Opus 5 91.8%
Android Bench 2.0 4 leaders
Claude Sonnet 5.5 40%

Source: Official leaderboards

Replaces AndroidBench

More in Coding

All 282 in Coding
  • SWE-bench Verified In use 500 human-validated GitHub issues from 12 Python repos. Agent writes a patch that must pass hidden tests. Claude Opus 4.7 83.5%
  • LiveCodeBench v6 In use Competitive programming problems collected after model cutoffs from LeetCode, AtCoder, Codeforces. Scored pass@1 on hidden tests. DeepSeek-V4-Pro 93.5%
  • SWE-bench Multilingual In use 300 GitHub-issue tasks across nine programming languages in SWE-bench format. Patch graded by repository tests. Claude Fable 5.2 96.6%
  • DeepSWE In use 113 original long-horizon engineering tasks across 91 repos and five languages, written from scratch. Hand-written verifiers accept alternative solutions. GPT-6 Astra 74.1%
  • SciCode In use Scientist-curated research coding problems split into subproblems across natural sciences. Code checked against test cases. Seed2.1 Pro 59.8%
  • Codeforces rating In use Model solves recent Codeforces contest problems. An Elo-style rating is computed from simulated contest performance. DeepSeek V4.1 Flash 3,471