DeepSeek V4 Pro (Workers AI) vs Gemini 3.1 Pro
DeepSeek V4 Pro (Workers AI) reads a 26.2K-token pull request for $0.0346, 34% less than Gemini 3.1 Pro ($0.0524). Both sit in the mid-range price tier, so the choice between them is about review quality more than budget. The benchmark cannot separate their recall: DeepSeek V4 Pro (Workers AI) minus Gemini 3.1 Pro is −13 pts (95% interval −26 pts to −1 pt).
Side by side
| Figure | DeepSeek V4 Pro (Workers AI) | Gemini 3.1 Pro |
|---|---|---|
| Price tier | Mid-range | Mid-range |
| Bugs caughtshare of known bugs | 54% | 67% (better) |
| False alarmsper clean PR | 2.9 | 0.8 (better) |
| Cost per reviewmeasured | $0.0643 (better) | $0.2499 |
| Cost per bug caught | $0.0872 (better) | $0.4134 |
| Exact linecatches on the expected lines | 89% | 89% (better) |
| Review timemean | 390s | 349s (better) |
| Typical review calllist price | $0.0442 (better) | $0.0814 |
| Input / 1M tokens | $1.32 (better) | $2.00 |
| Output / 1M tokens | $3.96 (better) | $12.00 |
| Context window | 1M | 1M |
Head to head
The known bugs both were scored on, by who caught each in at least one run.
By language
| Model | All | TypeScript6 bugs | Go4 bugs | Java4 bugs | Rust4 bugs | C#3 bugs | JavaScript3 bugs | Kotlin3 bugs | PHP3 bugs | Python3 bugs | Ruby3 bugs | Swift3 bugs | C++2 bugs | C1 bug |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| DeepSeek V4 Pro (Workers AI) | 54% | 79% | 25% | 49% | 54% | 59% | 17% | 30% | 69% | 72% | 67% | 75% | 38% | 33% |
| Gemini 3.1 Pro | 67% | 67% | 38% | 78% | 88% | 56% | 33% | 67% | 67% | 67% | 100% | 100% | 50% | 50% |
Of the 42 known bugs both were scored on, both caught 29 and neither caught 5. DeepSeek V4 Pro (Workers AI) caught 5 bugs that Gemini 3.1 Pro missed: 1 in C++, 1 in JavaScript, 1 in Kotlin, 1 in PHP and 1 in TypeScript, most of them security bugs. Gemini 3.1 Pro caught 3 bugs that DeepSeek V4 Pro (Workers AI) missed: 1 in Go, 1 in JavaScript and 1 in Rust, most of them concurrency bugs.
On pull requests with no known bug, DeepSeek V4 Pro (Workers AI) raised 2.9 false alarms per review and Gemini 3.1 Pro 0.8. A measured review cost $0.0643 on DeepSeek V4 Pro (Workers AI), 74% less than $0.2499 on Gemini 3.1 Pro.
| Pull request | Language · kind | DeepSeek V4 Pro (Workers AI) | Gemini 3.1 Pro |
|---|---|---|---|
| TanStack/queryfix(query-core): fix hydration bugs for already resolved promises | TypeScript · Logic | 3/4 | 0/1 |
| kotest/kotestFix DataClassEq comparator | Kotlin · Logic | 2/4 | 0/1 |
| nodejs/undicifeat: add configurable maxPayloadSize for WebSocket | JavaScript · Security | 2/4 | 0/1 |
| symfony/symfony[Security] Allow multiple OIDC discovery endpoints | PHP · Security | 1/3 | 0/1 |
| ggml-org/llama.cpprpc : cache and reuse compute graphs | C++ · Security | 1/4 | 0/1 |
| denoland/denoperf(libs/core): poll only ready handles and yield between I/O batches | Rust · Concurrency | 0/4 | 2/2 |
| golang/gocrypto/tls: add LocalCertificate to ConnectionState | Go · Logic | 0/4 | 1/2 |
| pinojs/thread-streamrefactor: replace setTimeout with Atomics.waitAsync | JavaScript · Concurrency | 0/4 | 1/1 |
Defects caught over all runs: a PR reviewed three times with one known bug shows 3/3 when it was caught every time.
Every bug each was scored on: DeepSeek V4 Pro (Workers AI) and Gemini 3.1 Pro.