✦ Our pick · Debug, fix & test
GPT-6.1 Sol at medium effort, $10.59 per finished task, is the cheapest you can get today. Within our measurement error it ties with Claude Sonnet 5.5 and Claude Opus 5.5: equally good bets, worth trying two.
Updated Oct 13, 2026, every 3 hours. See it live with your plans and costs.
The top 10 for debug, fix & test
Reproduce, locate and fix bugs; write and repair tests. Cost per finished task: the tokens at each model's best effort level for this work, plus your time fixing failed attempts ($93 an hour). Success is the share of such tasks a model gets right the first time, on the benchmarks below.
| # | Model | Effort | Success | $/task |
|---|---|---|---|---|
| 1 | Gemini 4 Argon Google | high | 79.3% partly measured | $10.19 |
| 2 | GPT-6.1 Sol OpenAI | medium | 76.8% partly measured | $10.59 |
| 3 | Claude Sonnet 5.5 Anthropic | xhigh | 78.7% partly measured | $10.90 |
| 4 | Claude Opus 5.5 Anthropic | medium | 77.7% partly measured | $11.41 |
| 5 | Claude Opus 5 Anthropic | medium | 76.3% measured | $13.65 |
| 6 | GPT-6 Sol OpenAI | high | 72.8% partly measured | $13.87 |
| 7 | MiMo-V2.6-Pro Xiaomi | default | 69.3% partly measured | $13.96 |
| 8 | GLM-5.3 Z AI | max | 69.7% partly measured | $14.52 |
| 9 | Ling 3.1 Flash InclusionAI | default | 67.4% partly measured | $15.27 |
| 10 | Claude Fable 5.1 Anthropic | high | 77.2% measured | $15.44 |
Best within your tool
| Tool | Model | Effort | $/task |
|---|---|---|---|
| Cursor | Claude Sonnet 5.5 | xhigh | $10.90 |
| Claude Code | Claude Sonnet 5.5 | xhigh | $10.90 |
| Codex | GPT-6.1 Sol | medium | $10.59 |
| Antigravity | Claude Sonnet 5.5 | xhigh | $10.90 |
| GitHub Copilot | GPT-6.1 Sol | medium | $10.59 |
If tokens come with your plan
Tokens included in a plan cost you nothing extra, so the pick is the most reliable model in it, at the leanest effort level within our error of the best (the top levels only use up the plan sooner).
| Plan | Model | Effort |
|---|---|---|
| Claude Pro | Claude Sonnet 5.5 | xhigh |
| Claude Max | Claude Sonnet 5.5 | xhigh |
| ChatGPT Plus or Business | GPT-6.1 Sol | medium |
| ChatGPT Pro | GPT-6.1 Sol | medium |
| Google AI Pro | Claude Sonnet 5.5 | xhigh |
| Google AI Ultra | Claude Sonnet 5.5 | xhigh |
Measured on
- DeepSWE v1.1 · Datacurve; models it hasn't run filled in from Artificial Analysis (agent runs). 113 original tasks across 91 real repositories in 5 languages, each requiring a large, novel change (not taken from public commits); pass@1 on hidden tests. Measured for 36 of the 42 ranked models, results as of Oct 13, 2026.
- Terminal-Bench 4.0 · Vals AI; models it hasn't run filled in from Artificial Analysis. 66 hard tasks in a real terminal: system administration, builds, environments, debugging, data processing. Measured for 39 of the 42 ranked models, results as of Oct 13, 2026.
- SWE Atlas Test Writing · Scale AI. 90 tasks: write production-grade tests for real codebases. Measured for 11 of the 42 ranked models.
How every number is worked out, with the errors of our estimates: the method.