Gemma 4 26B A4B
Google · Open weights · Released Apr 3, 2026 · google/gemma-4-26b-a4b-it
Best $0.05 in / $0.19 out via Darkbloom · int4
262K contextVisionToolsReasoningStructured outputPrompt cachingOpen weights
Available from 15 providersUSD / MTOK · PRICES 15 MINUTES AGO · SPEED 2 MINUTES AGO
EST. MONTHLY COSTclick a row for detail · click headers to sort
| PROVIDER | INPUT | OUTPUT | CACHE | CONTEXT | QUANT | TTFT | TPS | SMOKE | UPTIME |
|---|---|---|---|---|---|---|---|---|---|
| $0.05 | $0.19 | $0.013 | 262K | int4 | 1.8s | 80 tok/s | 5/5 | 100.0% | |
| DEDekaLLM | $0.06 | $0.33 | $0.04 | 262K | bf16 | 1.3s | 77 tok/s | — | 97.8% |
| $0.067 | $0.29 | $0.039 | 262K | bf16 | 0.86s | 110 tok/s | — | 100.0% | |
| $0.068 | $0.225 | $0.037 | 262K | bf16 | 0.62s | 108 tok/s | 5/5 | 99.9% | |
| $0.07 | $0.24 | $0.046 | 131K | fp4 | 1.8s | 168 tok/s | — | 100.0% | |
| $0.07 | $0.34 | — | 262K | fp8 | 0.67s | 58 tok/s | 5/5 | 100.0% | |
| $0.089 | $0.268 | $0.045 | 262K | bf16 | 1.1s | 106 tok/s | — | 100.0% | |
| MAMakora | $0.09 | $0.28 | $0.049 | 256K | — | 0.67s | 190 tok/s | — | 100.0% |
| $0.10 | $0.30 | $0.05 | 256K | — | 0.78s | 37 tok/s | 5/5 | 94.0% | |
| $0.10 | $0.30 | $0.05 | 262K | bf16 | 0.56s | 78 tok/s | — | 99.9% | |
| $0.13 | $0.40 | — | 262K | bf16 | ✕ | ✕ | 4/5 | 88.6% | |
| $0.13 | $0.40 | $0.05 | 262K | bf16 | 0.93s | 144 tok/s | — | 100.0% | |
| $0.13 | $0.40 | $0.05 | 256K | bf16 | 0.94s | 201 tok/s | — | 100.0% | |
| $0.14 | $0.40 | $0.05 | 262K | fp8 | 17.5s | 73 tok/s | 5/5 | 99.6% | |
| $0.15 | $0.60 | — | 262K | — | 0.62s | 108 tok/s | — | 99.7% |
TTFT/TPS: median (P50) over trailing 504h · chat10k profile (~10K in / 200 out) · measured via OpenRouter routing · ✕ = probed, endpoint returned no measurable stream · SMOKE: of 5 programmatic capability probes passed (JSON schema, instruction following, tool call, long-context, benign compliance; retry-once) · screened via OpenRouter routing · v0.2 · methodology v1.1 · measured 2 minutes ago
SERVE THIS MODEL AND NOT LISTED? LIST YOUR INFERENCE →
Input price history · best listed price, observed
↑ 19% SINCE OCT 30, 2026
$0.042 · OCT 30, 2026$0.05 · TODAY
Pricing variants · USD / MTok in / out
STANDARD
$0.05 / $0.19
per MTok in / out
FREE
Free / Free
per MTok in / out
CACHE READ $0.013
Recent changesRSS ↗
Nov 21, 2026price increased · input0.04 → 0.05
Nov 21, 2026price increased · output0.15 → 0.19
Nov 17, 2026price increased · output0.143 → 0.15
Nov 16, 2026price increased · input0.027 → 0.04
Nov 16, 2026price increased · output0.08 → 0.143
Nov 15, 2026price decreased · input0.033 → 0.027
Nov 15, 2026price decreased · output0.1 → 0.08
Nov 14, 2026price increased · input0.011 → 0.033
About
Gemma 4 26B A4B IT is an instruction-tuned Mixture-of-Experts (MoE) model from Google DeepMind. Despite 25.2B total parameters, only 3.8B activate per token during inference — delivering near-31B quality at...
Specifications
Context window262,144
Max output235,929
Modalities inimage, text, video
Modalities outtext
LicenseOpen weights
ReleasedApr 3, 2026
TokenizerGemma
WeightsHugging Face ↗
Open weights
safetensors
Related
Embed badge
[](https://modelindex.ai/models/google/gemma-4-26b-a4b-it)Sources
OpenRouter API15 minutes ago
AGGREGATORHugging Face15 minutes ago
COMMUNITYModelIndex benchmarks2 minutes ago
MEASUREDVerified 15 minutes ago · list prices, not negotiated rates.