AI Model Capability Comparison
Agentic coding
Only models released within the past year are shown. Only benchmarks covering at least 3 of these models are shown. Unknown release dates are excluded; historical data is retained.
— means no published score for this candidate in this source, not zero.
| No. | Model | 10/11 recent-year models have scores | 7/11 recent-year models have scores |
|---|---|---|---|
| 1 |  Claude Opus 5.5 | 54.64medium · claude-codeclaude-opus-5-5_medium | 62.33max · proximusclaude-opus-5-5_max |
| 2 |  GPT-6 Astra | 53.26max · codexgpt-6-astra_max | 65.51max · proximusgpt-6-astra_max |
| 3 | GPT-6.1 Sol | 50.23medium · codexgpt-6.1-sol_medium | — |
| 4 | Claude Fable 5.1 | 50.91medium · claude-codeclaude-fable-5-1_medium | 56.29max · proximusclaude-fable-5-1_max |
| 5 |  Gemini 4 Argon | 47.55medium · chiselgemini-4-argon_medium | 54.96high · proximusgemini-4-argon_high |
| 6 | Claude Sonnet 5.5 | 52.09xhigh · claude-codeclaude-sonnet-5-5_xhigh | 61.89max · proximusclaude-sonnet-5-5_max |
| 7 | Claude Opus 5 | 53.38medium · claude-codeclaude-opus-5_max | 52.01max · proximusclaude-opus-5_max |
| 8 | GPT-6 Sol | 49.27max · codexgpt-6-sol_max | — |
| 9 | GPT-6 Luna | 42.42max · codexgpt-6-luna_max | — |
| 10 | Claude Haiku 5.5 | 46.36max · claude-codeclaude-haiku-5-5_max | 43.77max · proximusclaude-haiku-5-5_max |
| 11 | Claude Mythos Preview | — | — |
Epoch AI benchmarks
Source: Epoch AI, CC BY 4.0 ↗ · Updated 2026-10-15
FrontierCode · Updated 2026-10-07
FrontierSWE · Updated 2026-10-07
PandaNpc Updated Oct 21, 2026, 06:01 UTC
Source: Epoch AI, CC BY 4.0 · Vals.ai · Artificial Analysis (artificialanalysis.ai)
Column guide
- No.
- Display order only. Sorting changes this number; it is not a published capability rank.
- Model
- The evaluated model name and its provider icon. A dash in a score column means no published score is available here, not a zero.
- FrontierCodeFrontierCode programming
- Tests whether code changes meet maintainers' requirements for merging, combining functional checks with quality criteria. Evaluation methodology ↗
- FrontierSWEFrontierSWE software engineering
- Tests difficult engineering and technical research tasks. Scores reflect the published task evaluation, with version and agent setup affecting comparability. Evaluation methodology ↗
Read scores within the same benchmark, source and version. Higher values indicate stronger performance on the reported metric, but different columns do not share a common scale.