SShogo Bench

Gemma 4 26B A4B IT

Open Gemma instruction model for efficient chat and self-hosted deployments

Shogo Agent Index

Unranked

Not ranked yet: 0 agentic benchmarks measured, and we need at least 2 plus an overall capability score. It will be ranked as results are published.

Agentic coding
—
estimated
Long-horizon autonomy
—
estimated
Knowledge-work agents
—
estimated
Computer use
—
estimated
Security & context learning
—
estimated

Price

Input
$0.10/M
Output
$0.30/M
Cache read
$0.050/M
3:1 blend
$0.15/M
Agent-effective
$0.070/M

Cheapest host for agents: Darkbloom at $0.020/M.

Agent-effective = 20 input tokens per output token, 85% cache hits. Why.

Agentic benchmarks

BenchmarkScoreNormalisedIndex weightHarness / source
SWE-bench Verifiedest.8%—
Terminal-Bench 2.0est.8%—
DeepSWEest.6%—
FrontierSWEest.5%—
CursorBenchest.4%—
GSO-Benchest.4%—
METR Time Horizonsest.10%—
Vending-Bench 2est.6%—
PostTrainBenchest.4%—
APEX-Agentsest.8%—
GDPvalest.7%—
Remote Labor Indexest.5%—
DeepResearch Benchest.5%—
OSWorld 2.0est.6%—
OSWorldest.4%—
Cybenchest.5%—
CL-benchest.5%—

API hosts

Live prices and uptime for every host serving Gemma 4 26B A4B IT via OpenRouter.

HostInputOutputCache readAgent $/MUptime 24hQuant
Darkbloom$0.040$0.15$0.010$0.020100.0%—
Io Net$0.067$0.29$0.039$0.050100.0%bf16
AkashML$0.070$0.24$0.046$0.06099.9%fp4
DekaLLM$0.060$0.33$0.040$0.06098.8%bf16
NextBit$0.076$0.26$0.043$0.06099.9%bf16
PrimeIntellect$0.080$0.24$0.040$0.060100.0%bf16
Cloudflare$0.10$0.30$0.050$0.07099.9%—
CoreWeave$0.10$0.30$0.050$0.070100.0%bf16
Makora$0.090$0.28$0.049$0.070100.0%—
DeepInfra$0.070$0.34—$0.080100.0%fp8
Parasail$0.13$0.40$0.050$0.080100.0%bf16
SiliconFlow$0.14$0.40$0.050$0.080100.0%fp8
Venice$0.13$0.40$0.050$0.080100.0%bf16
Novita$0.13$0.40—$0.1482.6%bf16
Google$0.15$0.60—$0.1784.3%—