Coding / Benchmark profile
An open benchmark for AI coding agents on real-world React Native implementation tasks, emphasizing working app behavior, recommended architecture choices, and strict constraint adherence.
Observed results
15 ranked models · higher is better · labels show rank and score
| Model | Rank | Model ID | Provider | Score | Unit |
|---|---|---|---|---|---|
| Composer 2 | #1 | composer-2 | Cursor | 96.1 | elo |
| Composer 2 Fast | #2 | composer-2-fast | Cursor | 94.9 | elo |
| GPT-5.4 | #3 | gpt-5-4 | OpenAI | 85.3 | elo |
| GPT-5.5 | #4 | gpt-5-5 | OpenAI | 84.7 | elo |
| Claude Opus 4.6 | #5 | claude-opus-4-6 | Anthropic | 84.1 | elo |
| Claude Opus 4.7 | #6 | claude-opus-4-7 | Anthropic | 82.8 | elo |
| Claude Sonnet 4.6 | #7 | claude-sonnet-4-6 | Anthropic | 80.6 | elo |
| Kimi K2.5 | #8 | kimi-k2-5 | Moonshot AI | 77.2 | elo |
| Gemma 4 31B | #9 | gemma-4-31b | 75.2 | elo | |
| GLM-5 | #10 | glm-5 | Z.AI | 74.8 | elo |
| Grok 4 | #11 | grok-4 | xAI | 72.6 | elo |
| GPT-OSS 120B | #12 | gpt-oss-120b | OpenAI | 71.6 | elo |
| DeepSeek V3.2 | #13 | deepseek-v3-2 | DeepSeek | 71.5 | elo |
| MiniMax M2.7 | #14 | minimax-m2-7 | MiniMax | 71.4 | elo |
| GPT-OSS 20B | #15 | gpt-oss-20b | OpenAI | 71 | elo |
One best compatible score per canonical product · higher is better
| Rank | Model | Weights | Evidence | Score |
|---|---|---|---|---|
| 01 | Composer 2Cursor | Closed weights | Source carrier | 96.1 elo |
| 02 | Composer 2 FastCursor | Closed weights | Source carrier | 94.9 elo |
| 03 | GPT-5.4OpenAI | Closed weights | Source carrier | 85.3 elo |
| 04 | GPT-5.5OpenAI | Closed weights | Source carrier | 84.7 elo |
| 05 | Claude Opus 4.6Anthropic | Closed weights | Source carrier | 84.1 elo |
| 06 | Claude Opus 4.7Anthropic | Closed weights | Source carrier | 82.8 elo |
| 07 | Claude Sonnet 4.6Anthropic | Closed weights | Source carrier | 80.6 elo |
| 08 | Kimi K2.5Moonshot AI | Open weights | Source carrier | 77.2 elo |
| 09 | Gemma 4 31BGoogle | Open weights | Source carrier | 75.2 elo |
| 10 | GLM-5Z.AI | Open weights | Source carrier | 74.8 elo |
Each distinct published measurement is retained. Same-snapshot canonical and configuration projections appear once.
| Published model / configuration | Score | Evidence & protocol | Source & dates |
|---|---|---|---|
Exact BenchLM registry variant Composer 2; bulk export does not retain a complete upstream harness configuration.Composer 2Exact identitySource label without a registered configuration ID Canonical product: composer-2 | 96.1 elo | Source carriersource-checkedVersion & system2026 Source-native system | BenchLM public datasets — 2026-08-01Observed Checked |
Exact BenchLM registry variant Composer 2 Fast; bulk export does not retain a complete upstream harness configuration.Composer 2 FastExact identitySource label without a registered configuration ID Canonical product: composer-2-fast | 94.9 elo | Source carriersource-checkedVersion & system2026 Source-native system | BenchLM public datasets — 2026-08-01Observed Checked |
Exact BenchLM registry variant GPT-5.4; bulk export does not retain a complete upstream harness configuration.GPT-5.4Exact identitySource label without a registered configuration ID Canonical product: gpt-5-4 | 85.3 elo | Source carriersource-checkedVersion & system2026 Source-native system | BenchLM public datasets — 2026-08-01Observed Checked |
Exact BenchLM registry variant GPT-5.5; bulk export does not retain a complete upstream harness configuration.GPT-5.5Exact identitySource label without a registered configuration ID Canonical product: gpt-5-5 | 84.7 elo | Source carriersource-checkedVersion & system2026 Source-native system | BenchLM public datasets — 2026-08-01Observed Checked |
Exact BenchLM registry variant Claude Opus 4.6; bulk export does not retain a complete upstream harness configuration.Claude Opus 4.6Exact identitySource label without a registered configuration ID Canonical product: claude-opus-4-6 | 84.1 elo | Source carriersource-checkedVersion & system2026 Source-native system | BenchLM public datasets — 2026-08-01Observed Checked |
Exact BenchLM registry variant Claude Opus 4.7; bulk export does not retain a complete upstream harness configuration.Claude Opus 4.7Exact identitySource label without a registered configuration ID Canonical product: claude-opus-4-7 | 82.8 elo | Source carriersource-checkedVersion & system2026 Source-native system | BenchLM public datasets — 2026-08-01Observed Checked |
Exact BenchLM registry variant Claude Sonnet 4.6; bulk export does not retain a complete upstream harness configuration.Claude Sonnet 4.6Exact identitySource label without a registered configuration ID Canonical product: claude-sonnet-4-6 | 80.6 elo | Source carriersource-checkedVersion & system2026 Source-native system | BenchLM public datasets — 2026-08-01Observed Checked |
Exact BenchLM registry variant Kimi K2.5; bulk export does not retain a complete upstream harness configuration.Kimi K2.5Exact identitySource label without a registered configuration ID Canonical product: kimi-k2-5 | 77.2 elo | Source carriersource-checkedVersion & system2026 Source-native system | BenchLM public datasets — 2026-08-01Observed Checked |
Exact BenchLM registry variant Gemma 4 31B; bulk export does not retain a complete upstream harness configuration.Gemma 4 31BExact identitySource label without a registered configuration ID Canonical product: gemma-4-31b | 75.2 elo | Source carriersource-checkedVersion & system2026 Source-native system | BenchLM public datasets — 2026-08-01Observed Checked |
Exact BenchLM registry variant GLM-5; bulk export does not retain a complete upstream harness configuration.GLM-5Exact identitySource label without a registered configuration ID Canonical product: glm-5 | 74.8 elo | Source carriersource-checkedVersion & system2026 Source-native system | BenchLM public datasets — 2026-08-01Observed Checked |
From result to context
An open benchmark for AI coding agents on real-world React Native implementation tasks, emphasizing working app behavior, recommended architecture choices, and strict constraint adherence.
The summary shows one best compatible source result per canonical product. Versions, effort settings and execution systems remain attached to the underlying records.
Scores from different versions or harnesses may not be interchangeable. The published source rows preserve those distinctions and their original units. A source result is not automatically an input to a current index.
This catalogue entry preserves available source evidence. Current indices admit only their specifically reviewed tracks and configurations.
Contamination risk: Unknown. Lifecycle: Active.
