Benchmark updates
Recent capture activity across benchmarked coding agents. Each row shows the latest version hibench recorded, when it was captured, and how default token footprint changed versus the previous version.
Benchmark data last updated June 24, 2026 UTC
| Agent | Version | Captured | Tokens (claude-opus-4-8) | Δ tokens | Tools | Skills |
|---|---|---|---|---|---|---|
| | v0.80.2 was v0.80.1 | June 24, 2026 | 1,977 | 0 + 0% | 4 0 | 0 0 |
Kilo Code | v7.3.54 was v7.3.53 | June 24, 2026 | 14,780 | 0 + 0% | 13 0 | 1 0 |
| | v0.2.64 was v0.2.63 | June 24, 2026 | 24,838 | 0 + 0% | 25 0 | 7 0 |
| | v1.0.64 was v1.0.63 | June 24, 2026 | 18,758 | -661 -3% | 16 -1 | 1 0 |
| | v0.157.1 was v0.157.0 | June 24, 2026 | 21,006 | +1 + 0% | 10 0 | 20 0 |
Devin | v2026.8.18 was v2026.7.23 | June 24, 2026 | 17,605 | -1,952 -10% | 23 0 | 2 0 |
| | v2026.06.24-00-45-58-9f61de7 was v2026.06.19-20-24-33-653a7fb | June 24, 2026 | 25,472 | +1 + 0% | 15 0 | 0 0 |
| | v0.142.0 was v0.141.0 | June 24, 2026 | 13,506 | 0 + 0% | 11 0 | 5 0 |
| | v2.1.187 was v2.1.186 | June 24, 2026 | 30,337 | -3,884 -11% | 24 -3 | 13 0 |
| | v2026.6.10 was v2026.6.9 | June 24, 2026 | 31,450 | -3 -0% | 33 0 | 0 0 |
Gemini CLI | v0.47.0 was v0.46.0 | June 22, 2026 | 13,789 | 0 + 0% | 8 0 | 2 0 |
| | v2.17.1 was v2.17.0 | June 21, 2026 | 12,393 | -1 -0% | 9 0 | 1 0 |
Cline | v3.0.29 was v3.0.28 | June 21, 2026 | 6,848 | 0 + 0% | 25 0 | 0 0 |
OpenCode | v1.17.9 was v1.17.8 | June 21, 2026 | 10,420 | 0 + 0% | 9 0 | 1 0 |
Hermes Agent | v0.17.0 was v0.16.0 | June 20, 2026 | 20,804 | +2,210 + 12% | 27 0 | 0 0 |
| | v1.16.0 was v1.15.0 | June 19, 2026 | 18,803 | -1,596 -8% | 7 +1 | 49 0 |
Sorted by most recent capture date. Token deltas compare the latest version to the immediately prior captured version using Anthropic claude-opus-4-8 totals.