GBA-Eval
We show this table for reference; we do not rank on it.
An agentic coding benchmark that asks models to build a Game Boy Advance emulator from scratch and grades emulator behavior against procedural, audio, and gameplay tests.
Overall score on GBA-Eval — 2026-07-26
We mirror the published overall score view for GBA-Eval. Claude Opus 5 leads the public snapshot at 79.6%, followed by Claude Fable 5 (74.5%) and Claude Opus 4.8 (70.9%). We do not use these results to rank models overall.
Claude Opus 5
Anthropic
Claude Fable 5
Anthropic
Claude Opus 4.8
Anthropic
23 modelsCodingCurrentDisplay onlyUpdated 2026-07-26
Overall score table (23 models)
| Rank | Model / configuration | Score | Parameters (B) | Open / closed |
|---|---|---|---|---|
| 1 | Claude Opus 5Anthropic | 79.6% | Not reported | Closed |
| 2 | Claude Fable 5Anthropic | 74.5% | Not reported | Closed |
| 3 | Claude Opus 4.8Anthropic | 70.9% | Not reported | Closed |
| 4 | Grok 4.5xAI | 65.4% | Not reported | Closed |
| 5 | Claude Sonnet 5Anthropic | 65.3% | Not reported | Closed |
| 6 | GPT-5.5OpenAI | 53.2% | Not reported | Closed |
| 7 | GPT-5.6 SolOpenAI | 52.6% | Not reported | Closed |
| 8 | Claude Sonnet 4.6Anthropic | 48.8% | Not reported | Closed |
| 9 | Kimi K3Moonshot AI | 48.3% | Not reported | Open |
| 10 | GPT-5.4OpenAI | 45.1% | Not reported | Closed |
| 11 | Claude Opus 4.6Anthropic | 44.1% | Not reported | Closed |
| 12 | Claude Opus 4.7Anthropic | 43.8% | Not reported | Closed |
| 13 | Muse Spark 1.1Meta | 7.9% | Not reported | Closed |
| 14 | Gemini 3.5 FlashGoogle | 6.7% | Not reported | Closed |
| 15 | Grok Build 0.1xAI | 2.4% | Not reported | Closed |
| 16 | MiniMax M3MiniMax | 0.9% | Not reported | Open |
| 17 | Kimi K2.6Moonshot AI | 0.9% | Not reported | Open |
| 18 | Kimi K2.7 CodeMoonshot AI | 0.9% | Not reported | Open |
| 19 | Gemini 3.1 ProGoogle | 0.8% | Not reported | Closed |
| 20 | Qwen3.7 MaxAlibaba | 0.4% | Not reported | Closed |
| 21 | GLM-5.2Z.AI | 0.0% | Not reported | Open |
| 22 | GLM-5.1Z.AI | 0.0% | Not reported | Open |
| 23 | MiniMax M2.7MiniMax | 0.0% | Not reported | Open |
How GBA-Eval is shown here
BenchLM mirrors the official GBA-Eval leaderboard snapshot graded on 2026-07-26. The benchmark asks coding agents to build a Game Boy Advance emulator and scores the result against 27 procedural, audio, and gameplay test cases.
GBA-Eval is display only on BenchLM. The source rows are agentic software-engineering runs with large token budgets and verifier-specific emulator tests, so BenchLM does not fold them into model-only weighted rankings.
Snapshot
The published GBA-Eval snapshot places Claude Opus 5 first at 79.6%. The third row is 8.7 points behind. The broader top-10 range is 34.5 points, so the table still separates the published systems.
23 models have been evaluated on GBA-Eval. The benchmark falls in the Coding category. We keep external benchmark mirrors separate from the weighted global scoring system, so these results remain source-specific evidence. GBA-Eval is currently displayed for reference but excluded from the scoring formula, so it does not directly affect overall rankings.
About GBA-Eval
Year
2026
Tasks
27 emulator test cases
Format
Overall emulator score
Difficulty
Long-horizon systems programming
GBA-Eval evaluates long-horizon coding agents by having them implement a working GBA emulator. The public leaderboard reports overall scores across 27 test cases with token usage and checkpoints preserved in the source JSON feed.
Freshness and provenance
Version
GBA-Eval 2026
Refresh cadence
Quarterly
Staleness state
Current
Question availability
Public benchmark set
BenchLM uses freshness metadata to decide whether a benchmark should still be treated as a strong differentiator, a benchmark to watch, or a display-only reference. For the full scoring policy, see the BenchLM methodology page.
Questions
What does GBA-Eval measure?
An agentic coding benchmark that asks models to build a Game Boy Advance emulator from scratch and grades emulator behavior against procedural, audio, and gameplay tests.
Which model leads the published GBA-Eval snapshot?
Claude Opus 5 currently leads the published GBA-Eval snapshot with 79.6% overall score. BenchLM shows this benchmark for display only and does not use it in overall rankings.
How many models are evaluated on GBA-Eval?
The 2026-07-26 snapshot contains 23 AI models.
Know when it’s worth switching models
The model to choose, the cheaper alternative, and the release we would wait on.
Read a sample issueJoin 7,000+ subscribers.
One email each week. Unsubscribe anytime.