GPT OSS 120B vs Gemma 4 26B A4B IT vs Qwen3 235B-A22B Instruct 2507
Gemma 4 26B A4B IT comes out ahead, 70 to 61 and 60 on our weighted score, and it is the cheaper option too.
Price in rating: Included · Exclude price
OpenAI
GPT OSS 120B
61/100- ECI140.2
- Price$0.15 / $0.60
- Context131K
- Our pick
Google
Gemma 4 26B A4B IT
70/100- ECI141.6
- Price$0.112 / $0.40
- Context262K
Alibaba (Qwen)
Qwen3 235B-A22B Instruct 2507
60/100- ECI138.1
- Price$0.15 / $0.60
- Context262K
Gemma 4 26B A4B IT is our pick
Gemma 4 26B A4B IT is the better all-round choice, scoring 70/100 against GPT OSS 120B (61) and Qwen3 235B-A22B Instruct 2507 (60). It leads on price and inputs & features. The score weighs capability 50%, price 25%, inputs & features 15%, context window 10%.
- CapabilityGemma 4 26B A4B ITCapabilities Index (ECI): Gemma 4 26B A4B IT 141.6 · GPT OSS 120B 140.2 · Qwen3 235B-A22B Instruct 2507 138.1
- Lowest priceGemma 4 26B A4B ITGemma 4 26B A4B IT $0.184 · GPT OSS 120B $0.263 · Qwen3 235B-A22B Instruct 2507 $0.263 per 1M tokens (3:1 blend)
- Longest contextGemma 4 26B A4B IT and Qwen3 235B-A22B Instruct 2507Gemma 4 26B A4B IT 262,144 · Qwen3 235B-A22B Instruct 2507 262,144 · GPT OSS 120B 131,072 tokens
- Widest inputsGemma 4 26B A4B ITGPT OSS 120B: Text · Gemma 4 26B A4B IT: Text, Images · Qwen3 235B-A22B Instruct 2507: Text
- Self-hostingAll open weightsEvery model here can be downloaded and run on your own hardware
| Measure | Weight | GPT OSS 120B | Gemma 4 26B A4B IT | Qwen3 235B-A22B Instruct 2507 |
|---|---|---|---|---|
| CapabilityCapabilities Index (ECI) | 50% | 68 | 69 | 66 |
| Price | 25% | 74 | 81 | 74 |
| Inputs & features | 15% | 35 | 70 | 25 |
| Context window | 10% | 32 | 43 | 43 |
| Overall | 100% | 61/100 | 70/100 | 60/100 |
Every spec in one table
Highlighted cells lead their row. Dashes mean the data is not published.
| Specification | |||
|---|---|---|---|
| Capability | |||
| Capabilities Index (ECI) | 140.2 | 141.6 (best) | 138.1 |
| ECI rank | #115 of 158 | #105 of 158 (best) | #120 of 158 |
| GPQA DiamondGraduate-level science questions | 75.8% (best) | 73.2% | — |
| OTIS Mock AIME 2024–2025Competition mathematics | 88.9% (best) | 82.2% | — |
| Price per million tokens | |||
| Input | $0.15 | $0.112 (best) | $0.15 |
| Output | $0.60 | $0.40 (best) | $0.60 |
| Cached input | — | — | — |
| Blended (3:1) | $0.263 | $0.184 (best) | $0.263 |
| Long-context rate | Same rate | Same rate | Same rate |
| Price source | Median of 77 providers | Median of 29 providers | Median of 23 providers |
| Limits | |||
| Context window | 131,072 tokens | 262,144 tokens (best) | 262,144 tokens (best) |
| Max output | 131,072 tokens (best) | 32,768 tokens | 16,384 tokens |
| Inputs and features | |||
| Text | Yes | Yes | Yes |
| Images | No | Yes | No |
| PDFs | No | No | No |
| Audio | No | No | No |
| Video | No | No | No |
| Reasoning | Yes | Yes | No |
| Tool calling | Yes | Yes | Yes |
| Structured output | No | Yes | No |
| Availability | |||
| Weights | Open | Open | OpenApache 2.0 |
| API model ID | — | gemma-4-26b-a4b-it | — |
| API providers | 80 (best) | 33 | 25 |
| Released | Aug 5, 2025 | Apr 2, 2026 | Jul 21, 2025 |
| Knowledge cutoff | Jun 2024 | — | — |
What would a month cost?
Enter your expected volume in millions of tokens. List prices only; caching and batch discounts would lower these.
GPT OSS 120B$2.70
Gemma 4 26B A4B IT$1.92
Qwen3 235B-A22B Instruct 2507$2.70
Which should you choose?
Which is better: GPT OSS 120B, Gemma 4 26B A4B IT or Qwen3 235B-A22B Instruct 2507?
Gemma 4 26B A4B IT is the better all-round choice, scoring 70/100 against GPT OSS 120B (61) and Qwen3 235B-A22B Instruct 2507 (60). It leads on price and inputs & features. The score weighs capability 50%, price 25%, inputs & features 15%, context window 10%.
Which is cheaper, GPT OSS 120B, Gemma 4 26B A4B IT or Qwen3 235B-A22B Instruct 2507?
Gemma 4 26B A4B IT is cheaper at $0.112 input / $0.40 output per million tokens (median across 29 API providers; free on Google). GPT OSS 120B costs $0.15 input / $0.60 output per million tokens (median across 77 API providers); Qwen3 235B-A22B Instruct 2507 costs $0.15 input / $0.60 output per million tokens (median across 23 API providers). At a typical mix of three input tokens to one output token, that is $0.184 per million tokens for Gemma 4 26B A4B IT versus $0.263 for GPT OSS 120B (1.4× as much) and $0.263 for Qwen3 235B-A22B Instruct 2507 (1.4× as much).
Which scores higher on benchmarks?
Gemma 4 26B A4B IT scores higher on the Capabilities Index (ECI): Gemma 4 26B A4B IT 141.6 (#105 of 158), GPT OSS 120B 140.2 (#115 of 158) and Qwen3 235B-A22B Instruct 2507 138.1 (#120 of 158). The confidence ranges of the top two overlap (139.1–144.0 vs 138.3–142.1), so treat the gap as small.
Which is better for coding?
There are no published SWE-bench Verified results for GPT OSS 120B, Gemma 4 26B A4B IT and Qwen3 235B-A22B Instruct 2507 yet, so there is no like-for-like coding score. On overall capability, Gemma 4 26B A4B IT leads, which tends to carry over to coding, but test on your own codebase. All three support tool calling for agent workflows.
Which has the bigger context window?
Gemma 4 26B A4B IT and Qwen3 235B-A22B Instruct 2507 have the largest context windows (262,144 and 262,144 tokens), against 131,072 for GPT OSS 120B. Maximum output per response: GPT OSS 120B up to 131,072, Gemma 4 26B A4B IT up to 32,768, Qwen3 235B-A22B Instruct 2507 up to 16,384 tokens.
Which can read images, PDFs, audio or video?
GPT OSS 120B accepts text; Gemma 4 26B A4B IT accepts text and images; Qwen3 235B-A22B Instruct 2507 accepts text. Gemma 4 26B A4B IT handles the widest range of inputs.
Are any of these open source?
Yes, all three publish their weights (Apache 2.0), so you can self-host them.
Which is newer?
Gemma 4 26B A4B IT is the newest, released Apr 2, 2026. GPT OSS 120B came out Aug 5, 2025; Qwen3 235B-A22B Instruct 2507 came out Jul 21, 2025. Knowledge cutoff: GPT OSS 120B Jun 2024.
How do you decide the winner?
This rating uses capability (50%, independent benchmark results); price (25%, blended price per million tokens (3 input : 1 output), log scale); inputs & features (15%, image, PDF, audio and video input, tool calling, structured output and reasoning); context window (10%, maximum tokens per request, log scale). Dimensions missing for any model are dropped and the remaining weights rescaled, so every model is judged on the same evidence. Specs and prices come from public model listings and the labs’ own API pages; capability scores come from independent benchmark runs. Data updated Nov 18, 2026.