Benchmark Almanac

GameXpert-Bench

Claude Opus 5 70.9 leads since August 2026
Not in model cards

Agents build games from 97 briefs, repair bugs in 100 tasks, and improve 17 games over six requests each. Each track reports its own scores.

Tencent compares 15 models per track. People and automated checks grade game creation. Chromium tests grade repairs. The improvement rubric checks the final game.

Released
August 2026
Kind
Building whole projects
Level
University
Scored
Graded by people
Answers
Agent task
Tasks
Kept private
Measures
Capability
Language
English
In and out
Text, Code → Code, Actions
Interaction
Agentic
Environment
Terminal, Browser
Refresh
One fixed set
Run by
A host
In views
Agents
Strongest model tested
Claude Opus 5 #11 on the capability index
Last model added
July 2026
Scores refreshed
November 20th, 2026
15
Models scored on its own leaderboard

Claude Opus 5 leads with 70.9 among models released by July 2026

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.

The newest model with a score here came out in July 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 5 GameFix: Claude Code, max 70.9
  2. 2 Claude Fable 5 GameFix: Claude Code, max 66.3
  3. 3 GPT-5.6 Sol GameFix: Claude Code, xhigh 57.9
  4. 4 Claude Opus 4.8 GameFix: Claude Code, max 56.8
  5. 5 Kimi K3 GameFix: Claude Code, max 56.6
  6. 6 Claude Opus 4.7 GameFix: Claude Code, max 53.9
  7. 7 DeepSeek-V4-Flash GameFix: Claude Code, max 52.1
  8. 8 GLM-5.2 GameFix: Claude Code, xhigh 49.9
  9. 9 GPT-5.5 GameFix: Claude Code, xhigh 49.8
  10. 10 Tencent Hy3 GameFix: Claude Code, high 45.7
  11. 11 MiniMax-M3 GameFix: Claude Code, on 43.3
  12. 12 GLM-5.1 GameFix: Claude Code, xhigh 41.9
  13. 13 Qwen3.7-Max GameFix: Claude Code, max 41.6
  14. 14 Gemini 3.5 Flash GameFix: Claude Code, high 39.5
  15. 15 Seed2.1 Pro GameFix: Claude Code, high 30.9

Source: Official GameXpert-Bench leaderboard

#ModelScoreSettingDateSource
1 Claude Opus 5 70.9 GameFix: Claude Code, max Jul 2026 Official leaderboard
2 Claude Fable 5 66.3 GameFix: Claude Code, max Jun 2026 Official leaderboard
3 GPT-5.6 Sol 57.9 GameFix: Claude Code, xhigh Jul 2026 Official leaderboard
4 Claude Opus 4.8 56.8 GameFix: Claude Code, max May 2026 Official leaderboard
5 Kimi K3 56.6 GameFix: Claude Code, max Jul 2026 Official leaderboard
6 Claude Opus 4.7 53.9 GameFix: Claude Code, max Apr 2026 Official leaderboard
7 DeepSeek-V4-Flash 52.1 GameFix: Claude Code, max Apr 2026 Official leaderboard
8 GLM-5.2 49.9 GameFix: Claude Code, xhigh Jun 2026 Official leaderboard
9 GPT-5.5 49.8 GameFix: Claude Code, xhigh Apr 2026 Official leaderboard
10 Tencent Hy3 45.7 GameFix: Claude Code, high Jul 2026 Official leaderboard
11 MiniMax-M3 43.3 GameFix: Claude Code, on Jun 2026 Official leaderboard
12 GLM-5.1 41.9 GameFix: Claude Code, xhigh Apr 2026 Official leaderboard
13 Qwen3.7-Max 41.6 GameFix: Claude Code, max May 2026 Official leaderboard
14 Gemini 3.5 Flash 39.5 GameFix: Claude Code, high May 2026 Official leaderboard
15 Seed2.1 Pro 30.9 GameFix: Claude Code, high Jun 2026 Official leaderboard

Rankings by part

We read 3 parts from the benchmark’s own leaderboard: GameGen, GameFix and GameOpt. Each ranks models on one part of the test. Claude Opus 5 leads every one of them. Arrows show how far a model moved from its overall rank among the models we track.

GameGen

15 models we track have a score here.

  1. 1 Claude Opus 5 79.7
  2. 2 Claude Fable 5 75.8
  3. 3 Kimi K3 ↑2 71.3
  4. 4 Claude Opus 4.8 69.9
  5. 5 GLM-5.2 ↑3 63.4
  6. 6 GPT-5.6 Sol ↓3 63.4
  7. 7 DeepSeek-V4-Flash 62.9
  8. 8 Tencent Hy3 ↑2 62.1
  9. 9 Claude Opus 4.7 ↓3 62.0
  10. 10 MiniMax-M3 ↑1 59.6

Number one over time

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

Anthropic
Own leaderboard Not tested since Aug 2026

Source: Official GameXpert-Bench 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
30.0 40.0 50.0 60.0 70.0 80.0 2026 30.0 40.0 50.0 60.0 70.0 80.0 2026

Source: Official GameXpert-Bench 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 Coding

All 523 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

Data audit

All 9 checks pass. Every benchmark

Catalog record 2 pass
  • Pass Task access The tasks are private.
  • Pass Confidence Checked against a primary source.
Scores 3 pass 2 notes
  • Pass Independent runs Scores from the official leaderboard (15 entries).
  • Pass Models scored 15 models we track.
  • Pass Newest score The newest is from July 2026, for Claude Opus 5.
  • Note Frontier models None of the top 10 has a score. The strongest tested is Claude Opus 5 (#11).
  • Note Headroom Scored as a plain number, with no fixed ceiling.
Leaderboard import 3 pass
  • Pass Importer We read the board on kwen-chen.github.io: 15 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