Alternatives to Gemma 4 31B

These models are alternatives by measurable coding evidence: they share the current closed-model classificationand are closest to Gemma 4 31B's SWE-bench score. Similarity here does not mean identical capabilities, licence terms, or output quality.

The target model

SWE-bench
1363.6%
Cheapest input
$0.08
Cheapest output
$0.34
Hosts
33
AlternativeSWE-benchInput $/MTokOutput $/MTokHostsContextTradeoff
Grok 4.201373.6%$1.25$2.5022.0Mhigher coding score, higher input price
Laguna M.11347.3%$0.20$0.401262khigher input price
GLM 4.61340.2%$0.39$1.7513198khigher input price
GPT-5.1-Codex1336.3%$1.25$10.003400khigher input price
MiMo-V2-Flash1330.2%$0.09$0.294262khigher input price
Inkling Small1401.6%$0.45$1.206524khigher coding score, higher input price
Inkling1407.6%$0.95$4.056524khigher coding score, higher input price
Laguna XS.21302.0%$0.10$0.201262khigher input price
MiniMax M21297.0%$0.26$1.007197khigher input price
MiMo-V2-Pro1433.5%$1.00$3.0011.0Mhigher coding score, higher input price
MiMo-V2.51437.8%$0.10$0.24131.1Mhigher coding score, higher input price
Gemini 3.1 Flash Lite1254.1%$0.13$0.7541.0Mhigher input price

How alternatives are selected

Alternatives must have a benchmark score and the same current open/closed classification as the target. They are sorted by score distance, then by price/performance. The table compares hosted API economics; it does not establish licence compatibility or self-hosting rights.

Check the model licence, provider Terms, context limits, tool support, and output quality before switching a production workload.