compare
OpenCode vs SWE-agent
The same facts for both, read from GitHub every night, and the relation a person reviewed.
| Fact | OpenCode | SWE-agent |
|---|---|---|
| Language | TypeScript | Python |
| Licence | MIT | MIT |
| Stars | 212k | 20k |
| Latest | v1.18.34 | v1.1.0 |
| Last push | 2026-10-06 | 2026-10-06 |
| Release cadence | about 6 days between releases | about 90 days between releases |
| Active contributors | 52+ commit authors on the default branch in the last 90 days | 2 commit authors on the default branch in the last 90 days |
| Flags | none | none |
How they relate
Both replace Claude Code. Alternatives to Claude Code →
Full replacementOpenCodeTerminal agent that works with Anthropic, OpenAI and Google models or a local one.
PartialSWE-agentTakes a GitHub issue and works on a fix with the model of your choice, aimed at autonomous runs rather than interactive sessions.
OpenCode
- v1.18.342026-09-30Send namespaced session and parent-session identity headers with model requests.
- v1.18.332026-09-28Cloudflare AI Gateway models now honor provider response and stream timeouts. (@danlapid)
- v1.18.322026-09-21Fixed Bedrock image attachments so they are only hoisted for Claude, Nova, and Llama 4 models.
- v1.18.312026-09-14Restored ACP session model, effort, mode, and reasoning chunk boundaries when loading, resuming, or forking sessions. (…
- v1.18.302026-09-09Added the Astra system prompt for GPT-6 models.
SWE-agent
- v1.1.02025-05-22v1.1.0: 10s of thousands of training trajectories
- v1.0.12025-02-28v1.0.1: SOTA on SWE-Bench Full
- v1.0.02025-02-13!swe_a_v1_announcement_v6
- v0.7.02024-09-25SWE-agent EnIGMA (0.7.0)
- v0.6.12024-06-20This is (mostly) a patch release, in particular fixing several issues that had been introduced by the speed improvement…
OpenCode
unsigned The latest release, v1.18.34, carries no signature GitHub could verify.
Loading the security report
SWE-agent
✓ signed The latest release, v1.1.0, carries a signature GitHub verified.
Loading the security report